{ "best_metric": 0.5833333333333334, "best_model_checkpoint": "/home/pricie/trusca/adapter_trocr/data/models_magister_dixit/ner_model/checkpoint-3400", "epoch": 242.85714285714286, "eval_steps": 200, "global_step": 3400, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 3.571428571428571, "grad_norm": 0.09315240383148193, "learning_rate": 4.966428571428572e-05, "loss": 0.1403, "step": 50 }, { "epoch": 7.142857142857143, "grad_norm": 0.07808659970760345, "learning_rate": 4.930714285714286e-05, "loss": 0.0135, "step": 100 }, { "epoch": 10.714285714285714, "grad_norm": 0.08603320270776749, "learning_rate": 4.8950000000000004e-05, "loss": 0.0062, "step": 150 }, { "epoch": 14.285714285714286, "grad_norm": 0.18650108575820923, "learning_rate": 4.8592857142857144e-05, "loss": 0.0033, "step": 200 }, { "epoch": 14.285714285714286, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6077348066298343, "number": 88, "precision": 0.5913978494623656, "recall": 0.625 }, "eval_loss": 0.027063699439167976, "eval_overall_accuracy": 0.9959282078757032, "eval_overall_f1": 0.5759162303664922, "eval_overall_precision": 0.5789473684210527, "eval_overall_recall": 0.5729166666666666, "eval_runtime": 0.3664, "eval_samples_per_second": 150.093, "eval_steps_per_second": 2.729, "step": 200 }, { "epoch": 17.857142857142858, "grad_norm": 0.07374002784490585, "learning_rate": 4.823571428571429e-05, "loss": 0.0019, "step": 250 }, { "epoch": 21.428571428571427, "grad_norm": 0.06003947556018829, "learning_rate": 4.787857142857143e-05, "loss": 0.0011, "step": 300 }, { "epoch": 25.0, "grad_norm": 0.039870865643024445, "learning_rate": 4.752142857142857e-05, "loss": 0.0008, "step": 350 }, { "epoch": 28.571428571428573, "grad_norm": 0.04143398255109787, "learning_rate": 4.716428571428572e-05, "loss": 0.0006, "step": 400 }, { "epoch": 28.571428571428573, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6179775280898877, "number": 88, "precision": 0.6111111111111112, "recall": 0.625 }, "eval_loss": 0.030486594885587692, "eval_overall_accuracy": 0.9958746316635414, "eval_overall_f1": 0.5729166666666666, "eval_overall_precision": 0.5729166666666666, "eval_overall_recall": 0.5729166666666666, "eval_runtime": 0.3637, "eval_samples_per_second": 151.235, "eval_steps_per_second": 2.75, "step": 400 }, { "epoch": 32.142857142857146, "grad_norm": 0.034276727586984634, "learning_rate": 4.680714285714286e-05, "loss": 0.0004, "step": 450 }, { "epoch": 35.714285714285715, "grad_norm": 0.08624289929866791, "learning_rate": 4.6450000000000004e-05, "loss": 0.0003, "step": 500 }, { "epoch": 39.285714285714285, "grad_norm": 0.026272911578416824, "learning_rate": 4.6092857142857144e-05, "loss": 0.0002, "step": 550 }, { "epoch": 42.857142857142854, "grad_norm": 0.02422301098704338, "learning_rate": 4.5735714285714284e-05, "loss": 0.0003, "step": 600 }, { "epoch": 42.857142857142854, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.5921787709497207, "number": 88, "precision": 0.5824175824175825, "recall": 0.6022727272727273 }, "eval_loss": 0.03127627447247505, "eval_overall_accuracy": 0.995713903027056, "eval_overall_f1": 0.5549738219895287, "eval_overall_precision": 0.5578947368421052, "eval_overall_recall": 0.5520833333333334, "eval_runtime": 0.3675, "eval_samples_per_second": 149.668, "eval_steps_per_second": 2.721, "step": 600 }, { "epoch": 46.42857142857143, "grad_norm": 0.026417318731546402, "learning_rate": 4.537857142857143e-05, "loss": 0.0003, "step": 650 }, { "epoch": 50.0, "grad_norm": 0.042232610285282135, "learning_rate": 4.502142857142858e-05, "loss": 0.0003, "step": 700 }, { "epoch": 53.57142857142857, "grad_norm": 0.03797277435660362, "learning_rate": 4.466428571428571e-05, "loss": 0.0002, "step": 750 }, { "epoch": 57.142857142857146, "grad_norm": 0.17620719969272614, "learning_rate": 4.430714285714286e-05, "loss": 0.0002, "step": 800 }, { "epoch": 57.142857142857146, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6265060240963856, "number": 88, "precision": 0.6666666666666666, "recall": 0.5909090909090909 }, "eval_loss": 0.03513283655047417, "eval_overall_accuracy": 0.9963032413608358, "eval_overall_f1": 0.5909090909090908, "eval_overall_precision": 0.65, "eval_overall_recall": 0.5416666666666666, "eval_runtime": 0.3731, "eval_samples_per_second": 147.431, "eval_steps_per_second": 2.681, "step": 800 }, { "epoch": 60.714285714285715, "grad_norm": 0.004675676114857197, "learning_rate": 4.3950000000000004e-05, "loss": 0.0003, "step": 850 }, { "epoch": 64.28571428571429, "grad_norm": 0.04627036675810814, "learning_rate": 4.3592857142857144e-05, "loss": 0.0003, "step": 900 }, { "epoch": 67.85714285714286, "grad_norm": 0.011826109141111374, "learning_rate": 4.323571428571429e-05, "loss": 0.0002, "step": 950 }, { "epoch": 71.42857142857143, "grad_norm": 0.006843605078756809, "learning_rate": 4.287857142857143e-05, "loss": 0.0003, "step": 1000 }, { "epoch": 71.42857142857143, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6432748538011697, "number": 88, "precision": 0.6626506024096386, "recall": 0.625 }, "eval_loss": 0.03261447325348854, "eval_overall_accuracy": 0.9964639699973212, "eval_overall_f1": 0.5978260869565217, "eval_overall_precision": 0.625, "eval_overall_recall": 0.5729166666666666, "eval_runtime": 0.3662, "eval_samples_per_second": 150.195, "eval_steps_per_second": 2.731, "step": 1000 }, { "epoch": 75.0, "grad_norm": 0.03144918382167816, "learning_rate": 4.252142857142857e-05, "loss": 0.0002, "step": 1050 }, { "epoch": 78.57142857142857, "grad_norm": 0.014924576506018639, "learning_rate": 4.216428571428572e-05, "loss": 0.0001, "step": 1100 }, { "epoch": 82.14285714285714, "grad_norm": 0.017610769718885422, "learning_rate": 4.1807142857142865e-05, "loss": 0.0002, "step": 1150 }, { "epoch": 85.71428571428571, "grad_norm": 0.0011982271680608392, "learning_rate": 4.145e-05, "loss": 0.0001, "step": 1200 }, { "epoch": 85.71428571428571, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6206896551724139, "number": 88, "precision": 0.627906976744186, "recall": 0.6136363636363636 }, "eval_loss": 0.034781765192747116, "eval_overall_accuracy": 0.9960889365121885, "eval_overall_f1": 0.5806451612903225, "eval_overall_precision": 0.6, "eval_overall_recall": 0.5625, "eval_runtime": 0.363, "eval_samples_per_second": 151.51, "eval_steps_per_second": 2.755, "step": 1200 }, { "epoch": 89.28571428571429, "grad_norm": 0.0015545673668384552, "learning_rate": 4.1092857142857145e-05, "loss": 0.0001, "step": 1250 }, { "epoch": 92.85714285714286, "grad_norm": 0.00382990506477654, "learning_rate": 4.073571428571429e-05, "loss": 0.0001, "step": 1300 }, { "epoch": 96.42857142857143, "grad_norm": 0.004094763193279505, "learning_rate": 4.037857142857143e-05, "loss": 0.0002, "step": 1350 }, { "epoch": 100.0, "grad_norm": 0.006723235361278057, "learning_rate": 4.002142857142857e-05, "loss": 0.0001, "step": 1400 }, { "epoch": 100.0, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.654320987654321, "number": 88, "precision": 0.7162162162162162, "recall": 0.6022727272727273 }, "eval_loss": 0.0324268601834774, "eval_overall_accuracy": 0.9966246986338065, "eval_overall_f1": 0.6091954022988507, "eval_overall_precision": 0.6794871794871795, "eval_overall_recall": 0.5520833333333334, "eval_runtime": 0.3666, "eval_samples_per_second": 150.021, "eval_steps_per_second": 2.728, "step": 1400 }, { "epoch": 103.57142857142857, "grad_norm": 0.021055925637483597, "learning_rate": 3.966428571428572e-05, "loss": 0.0001, "step": 1450 }, { "epoch": 107.14285714285714, "grad_norm": 0.03446558490395546, "learning_rate": 3.930714285714286e-05, "loss": 0.0002, "step": 1500 }, { "epoch": 110.71428571428571, "grad_norm": 0.0060461643151938915, "learning_rate": 3.8950000000000005e-05, "loss": 0.0001, "step": 1550 }, { "epoch": 114.28571428571429, "grad_norm": 0.025492200627923012, "learning_rate": 3.8592857142857145e-05, "loss": 0.0001, "step": 1600 }, { "epoch": 114.28571428571429, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6506024096385543, "number": 88, "precision": 0.6923076923076923, "recall": 0.6136363636363636 }, "eval_loss": 0.03459694981575012, "eval_overall_accuracy": 0.996517546209483, "eval_overall_f1": 0.6067415730337078, "eval_overall_precision": 0.6585365853658537, "eval_overall_recall": 0.5625, "eval_runtime": 0.3562, "eval_samples_per_second": 154.408, "eval_steps_per_second": 2.807, "step": 1600 }, { "epoch": 117.85714285714286, "grad_norm": 0.0011437517823651433, "learning_rate": 3.8235714285714285e-05, "loss": 0.0001, "step": 1650 }, { "epoch": 121.42857142857143, "grad_norm": 0.0015667779371142387, "learning_rate": 3.787857142857143e-05, "loss": 0.0002, "step": 1700 }, { "epoch": 125.0, "grad_norm": 0.0007876116433180869, "learning_rate": 3.752142857142857e-05, "loss": 0.0002, "step": 1750 }, { "epoch": 128.57142857142858, "grad_norm": 0.000237628526519984, "learning_rate": 3.716428571428572e-05, "loss": 0.0001, "step": 1800 }, { "epoch": 128.57142857142858, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6463414634146342, "number": 88, "precision": 0.6973684210526315, "recall": 0.6022727272727273 }, "eval_loss": 0.03679116442799568, "eval_overall_accuracy": 0.996517546209483, "eval_overall_f1": 0.6022727272727273, "eval_overall_precision": 0.6625, "eval_overall_recall": 0.5520833333333334, "eval_runtime": 0.3661, "eval_samples_per_second": 150.247, "eval_steps_per_second": 2.732, "step": 1800 }, { "epoch": 132.14285714285714, "grad_norm": 0.0020967507734894753, "learning_rate": 3.680714285714286e-05, "loss": 0.0001, "step": 1850 }, { "epoch": 135.71428571428572, "grad_norm": 0.0021254734601825476, "learning_rate": 3.645e-05, "loss": 0.0001, "step": 1900 }, { "epoch": 139.28571428571428, "grad_norm": 0.0005617139395326376, "learning_rate": 3.6092857142857145e-05, "loss": 0.0001, "step": 1950 }, { "epoch": 142.85714285714286, "grad_norm": 0.04763318970799446, "learning_rate": 3.573571428571429e-05, "loss": 0.0001, "step": 2000 }, { "epoch": 142.85714285714286, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6153846153846153, "number": 88, "precision": 0.6419753086419753, "recall": 0.5909090909090909 }, "eval_loss": 0.038500551134347916, "eval_overall_accuracy": 0.9961425127243504, "eval_overall_f1": 0.574585635359116, "eval_overall_precision": 0.611764705882353, "eval_overall_recall": 0.5416666666666666, "eval_runtime": 0.3649, "eval_samples_per_second": 150.735, "eval_steps_per_second": 2.741, "step": 2000 }, { "epoch": 146.42857142857142, "grad_norm": 0.007793730590492487, "learning_rate": 3.5378571428571425e-05, "loss": 0.0001, "step": 2050 }, { "epoch": 150.0, "grad_norm": 0.00020323636999819428, "learning_rate": 3.502142857142857e-05, "loss": 0.0001, "step": 2100 }, { "epoch": 153.57142857142858, "grad_norm": 0.004183015786111355, "learning_rate": 3.466428571428572e-05, "loss": 0.0001, "step": 2150 }, { "epoch": 157.14285714285714, "grad_norm": 0.04455306753516197, "learning_rate": 3.430714285714286e-05, "loss": 0.0001, "step": 2200 }, { "epoch": 157.14285714285714, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6415094339622642, "number": 88, "precision": 0.7183098591549296, "recall": 0.5795454545454546 }, "eval_loss": 0.035211097449064255, "eval_overall_accuracy": 0.9965711224216448, "eval_overall_f1": 0.5964912280701754, "eval_overall_precision": 0.68, "eval_overall_recall": 0.53125, "eval_runtime": 0.3613, "eval_samples_per_second": 152.208, "eval_steps_per_second": 2.767, "step": 2200 }, { "epoch": 160.71428571428572, "grad_norm": 0.03668719530105591, "learning_rate": 3.3950000000000005e-05, "loss": 0.0001, "step": 2250 }, { "epoch": 164.28571428571428, "grad_norm": 0.0001423813810106367, "learning_rate": 3.3592857142857145e-05, "loss": 0.0001, "step": 2300 }, { "epoch": 167.85714285714286, "grad_norm": 0.06529837846755981, "learning_rate": 3.3235714285714285e-05, "loss": 0.0001, "step": 2350 }, { "epoch": 171.42857142857142, "grad_norm": 0.018087288364768028, "learning_rate": 3.287857142857143e-05, "loss": 0.0001, "step": 2400 }, { "epoch": 171.42857142857142, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6227544910179641, "number": 88, "precision": 0.6582278481012658, "recall": 0.5909090909090909 }, "eval_loss": 0.03764405846595764, "eval_overall_accuracy": 0.996249665148674, "eval_overall_f1": 0.5810055865921787, "eval_overall_precision": 0.6265060240963856, "eval_overall_recall": 0.5416666666666666, "eval_runtime": 0.3612, "eval_samples_per_second": 152.256, "eval_steps_per_second": 2.768, "step": 2400 }, { "epoch": 175.0, "grad_norm": 0.0019430763786658645, "learning_rate": 3.252142857142857e-05, "loss": 0.0001, "step": 2450 }, { "epoch": 178.57142857142858, "grad_norm": 0.05669141560792923, "learning_rate": 3.216428571428571e-05, "loss": 0.0001, "step": 2500 }, { "epoch": 182.14285714285714, "grad_norm": 0.0005553144146688282, "learning_rate": 3.180714285714286e-05, "loss": 0.0001, "step": 2550 }, { "epoch": 185.71428571428572, "grad_norm": 0.024925246834754944, "learning_rate": 3.145e-05, "loss": 0.0002, "step": 2600 }, { "epoch": 185.71428571428572, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.5925925925925926, "number": 88, "precision": 0.6486486486486487, "recall": 0.5454545454545454 }, "eval_loss": 0.036129821091890335, "eval_overall_accuracy": 0.9960889365121885, "eval_overall_f1": 0.5517241379310345, "eval_overall_precision": 0.6153846153846154, "eval_overall_recall": 0.5, "eval_runtime": 0.3617, "eval_samples_per_second": 152.062, "eval_steps_per_second": 2.765, "step": 2600 }, { "epoch": 189.28571428571428, "grad_norm": 0.00033352759783156216, "learning_rate": 3.1092857142857145e-05, "loss": 0.0001, "step": 2650 }, { "epoch": 192.85714285714286, "grad_norm": 0.07385773211717606, "learning_rate": 3.073571428571429e-05, "loss": 0.0001, "step": 2700 }, { "epoch": 196.42857142857142, "grad_norm": 0.007033195812255144, "learning_rate": 3.037857142857143e-05, "loss": 0.0001, "step": 2750 }, { "epoch": 200.0, "grad_norm": 0.005518068093806505, "learning_rate": 3.0021428571428572e-05, "loss": 0.0001, "step": 2800 }, { "epoch": 200.0, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6190476190476191, "number": 88, "precision": 0.65, "recall": 0.5909090909090909 }, "eval_loss": 0.038286928087472916, "eval_overall_accuracy": 0.9960889365121885, "eval_overall_f1": 0.5714285714285714, "eval_overall_precision": 0.6046511627906976, "eval_overall_recall": 0.5416666666666666, "eval_runtime": 0.3616, "eval_samples_per_second": 152.119, "eval_steps_per_second": 2.766, "step": 2800 }, { "epoch": 203.57142857142858, "grad_norm": 0.0010896483436226845, "learning_rate": 2.9664285714285716e-05, "loss": 0.0001, "step": 2850 }, { "epoch": 207.14285714285714, "grad_norm": 0.00024828550522215664, "learning_rate": 2.9307142857142862e-05, "loss": 0.0002, "step": 2900 }, { "epoch": 210.71428571428572, "grad_norm": 0.001715456135571003, "learning_rate": 2.895e-05, "loss": 0.0001, "step": 2950 }, { "epoch": 214.28571428571428, "grad_norm": 0.029867948964238167, "learning_rate": 2.8592857142857142e-05, "loss": 0.0001, "step": 3000 }, { "epoch": 214.28571428571428, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6424242424242423, "number": 88, "precision": 0.6883116883116883, "recall": 0.6022727272727273 }, "eval_loss": 0.03977252170443535, "eval_overall_accuracy": 0.9964103937851594, "eval_overall_f1": 0.5955056179775281, "eval_overall_precision": 0.6463414634146342, "eval_overall_recall": 0.5520833333333334, "eval_runtime": 0.3638, "eval_samples_per_second": 151.198, "eval_steps_per_second": 2.749, "step": 3000 }, { "epoch": 217.85714285714286, "grad_norm": 0.0017272685654461384, "learning_rate": 2.823571428571429e-05, "loss": 0.0001, "step": 3050 }, { "epoch": 221.42857142857142, "grad_norm": 0.0022916370071470737, "learning_rate": 2.7878571428571432e-05, "loss": 0.0001, "step": 3100 }, { "epoch": 225.0, "grad_norm": 0.002857557963579893, "learning_rate": 2.752142857142857e-05, "loss": 0.0001, "step": 3150 }, { "epoch": 228.57142857142858, "grad_norm": 0.00026950574829243124, "learning_rate": 2.7164285714285716e-05, "loss": 0.0001, "step": 3200 }, { "epoch": 228.57142857142858, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6347305389221557, "number": 88, "precision": 0.6708860759493671, "recall": 0.6022727272727273 }, "eval_loss": 0.0408807098865509, "eval_overall_accuracy": 0.996249665148674, "eval_overall_f1": 0.585635359116022, "eval_overall_precision": 0.6235294117647059, "eval_overall_recall": 0.5520833333333334, "eval_runtime": 0.3628, "eval_samples_per_second": 151.609, "eval_steps_per_second": 2.757, "step": 3200 }, { "epoch": 232.14285714285714, "grad_norm": 5.138161577633582e-05, "learning_rate": 2.680714285714286e-05, "loss": 0.0001, "step": 3250 }, { "epoch": 235.71428571428572, "grad_norm": 0.024591689929366112, "learning_rate": 2.6450000000000003e-05, "loss": 0.0001, "step": 3300 }, { "epoch": 239.28571428571428, "grad_norm": 0.011367405764758587, "learning_rate": 2.6092857142857146e-05, "loss": 0.0001, "step": 3350 }, { "epoch": 242.85714285714286, "grad_norm": 0.0027745633851736784, "learning_rate": 2.5735714285714286e-05, "loss": 0.0001, "step": 3400 }, { "epoch": 242.85714285714286, "eval_LOC": { "f1": 0.0, "number": 8, "precision": 0.0, "recall": 0.0 }, "eval_PERS": { "f1": 0.6436781609195402, "number": 88, "precision": 0.6511627906976745, "recall": 0.6363636363636364 }, "eval_loss": 0.040758177638053894, "eval_overall_accuracy": 0.9963032413608358, "eval_overall_f1": 0.6021505376344085, "eval_overall_precision": 0.6222222222222222, "eval_overall_recall": 0.5833333333333334, "eval_runtime": 0.3617, "eval_samples_per_second": 152.076, "eval_steps_per_second": 2.765, "step": 3400 } ], "logging_steps": 50, "max_steps": 7000, "num_input_tokens_seen": 0, "num_train_epochs": 500, "save_steps": 200, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 5.62273611174912e+16, "train_batch_size": 64, "trial_name": null, "trial_params": null }