| { | |
| "best_metric": 0.19315651059150696, | |
| "best_model_checkpoint": "./vit-base-brain-alzheimer-detection/checkpoint-2800", | |
| "epoch": 13.28125, | |
| "eval_steps": 200, | |
| "global_step": 3400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1953125, | |
| "grad_norm": 3.182270050048828, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 1.401, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.390625, | |
| "grad_norm": 2.6782896518707275, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 1.1871, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5859375, | |
| "grad_norm": 2.7876715660095215, | |
| "learning_rate": 6e-06, | |
| "loss": 1.035, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.78125, | |
| "grad_norm": 3.130065441131592, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.9772, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.78125, | |
| "eval_accuracy": 0.580078125, | |
| "eval_loss": 0.9400424957275391, | |
| "eval_runtime": 5.935, | |
| "eval_samples_per_second": 172.535, | |
| "eval_steps_per_second": 21.567, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.9765625, | |
| "grad_norm": 2.5494868755340576, | |
| "learning_rate": 1e-05, | |
| "loss": 0.8886, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.171875, | |
| "grad_norm": 1.62320876121521, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.8722, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.3671875, | |
| "grad_norm": 4.327115535736084, | |
| "learning_rate": 1.4e-05, | |
| "loss": 0.8154, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.5625, | |
| "grad_norm": 7.814743995666504, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.7451, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.5625, | |
| "eval_accuracy": 0.6552734375, | |
| "eval_loss": 0.7946789264678955, | |
| "eval_runtime": 7.068, | |
| "eval_samples_per_second": 144.879, | |
| "eval_steps_per_second": 18.11, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.7578125, | |
| "grad_norm": 15.593164443969727, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.7383, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.953125, | |
| "grad_norm": 5.583272933959961, | |
| "learning_rate": 2e-05, | |
| "loss": 0.733, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.1484375, | |
| "grad_norm": 10.433585166931152, | |
| "learning_rate": 1.9783549783549785e-05, | |
| "loss": 0.5905, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.34375, | |
| "grad_norm": 4.751012802124023, | |
| "learning_rate": 1.9567099567099568e-05, | |
| "loss": 0.5701, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.34375, | |
| "eval_accuracy": 0.7236328125, | |
| "eval_loss": 0.7641921639442444, | |
| "eval_runtime": 5.5404, | |
| "eval_samples_per_second": 184.824, | |
| "eval_steps_per_second": 23.103, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.5390625, | |
| "grad_norm": 17.00213050842285, | |
| "learning_rate": 1.9350649350649354e-05, | |
| "loss": 0.6235, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.734375, | |
| "grad_norm": 10.933674812316895, | |
| "learning_rate": 1.9134199134199138e-05, | |
| "loss": 0.5385, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.9296875, | |
| "grad_norm": 9.459794044494629, | |
| "learning_rate": 1.891774891774892e-05, | |
| "loss": 0.4557, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 3.125, | |
| "grad_norm": 2.7128560543060303, | |
| "learning_rate": 1.8701298701298704e-05, | |
| "loss": 0.3704, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.125, | |
| "eval_accuracy": 0.7744140625, | |
| "eval_loss": 0.5531814098358154, | |
| "eval_runtime": 6.1892, | |
| "eval_samples_per_second": 165.45, | |
| "eval_steps_per_second": 20.681, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.3203125, | |
| "grad_norm": 6.759947776794434, | |
| "learning_rate": 1.8484848484848487e-05, | |
| "loss": 0.3876, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 3.515625, | |
| "grad_norm": 4.957170009613037, | |
| "learning_rate": 1.826839826839827e-05, | |
| "loss": 0.3014, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.7109375, | |
| "grad_norm": 4.334908485412598, | |
| "learning_rate": 1.8051948051948053e-05, | |
| "loss": 0.2523, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 3.90625, | |
| "grad_norm": 12.843859672546387, | |
| "learning_rate": 1.7835497835497836e-05, | |
| "loss": 0.2906, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.90625, | |
| "eval_accuracy": 0.85546875, | |
| "eval_loss": 0.4422951340675354, | |
| "eval_runtime": 5.5363, | |
| "eval_samples_per_second": 184.962, | |
| "eval_steps_per_second": 23.12, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.1015625, | |
| "grad_norm": 3.9840188026428223, | |
| "learning_rate": 1.761904761904762e-05, | |
| "loss": 0.2216, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 4.296875, | |
| "grad_norm": 11.730119705200195, | |
| "learning_rate": 1.7402597402597403e-05, | |
| "loss": 0.1722, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.4921875, | |
| "grad_norm": 2.089110851287842, | |
| "learning_rate": 1.718614718614719e-05, | |
| "loss": 0.1427, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 4.6875, | |
| "grad_norm": 0.5950905084609985, | |
| "learning_rate": 1.6969696969696972e-05, | |
| "loss": 0.1636, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.6875, | |
| "eval_accuracy": 0.900390625, | |
| "eval_loss": 0.32256966829299927, | |
| "eval_runtime": 6.2703, | |
| "eval_samples_per_second": 163.31, | |
| "eval_steps_per_second": 20.414, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.8828125, | |
| "grad_norm": 4.693203926086426, | |
| "learning_rate": 1.6753246753246756e-05, | |
| "loss": 0.1622, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 5.078125, | |
| "grad_norm": 0.24195541441440582, | |
| "learning_rate": 1.653679653679654e-05, | |
| "loss": 0.099, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 5.2734375, | |
| "grad_norm": 0.16812123358249664, | |
| "learning_rate": 1.6320346320346322e-05, | |
| "loss": 0.0979, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 5.46875, | |
| "grad_norm": 19.697050094604492, | |
| "learning_rate": 1.6103896103896105e-05, | |
| "loss": 0.0837, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.46875, | |
| "eval_accuracy": 0.90234375, | |
| "eval_loss": 0.3483480215072632, | |
| "eval_runtime": 5.5142, | |
| "eval_samples_per_second": 185.702, | |
| "eval_steps_per_second": 23.213, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.6640625, | |
| "grad_norm": 0.35744136571884155, | |
| "learning_rate": 1.5887445887445888e-05, | |
| "loss": 0.118, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 5.859375, | |
| "grad_norm": 0.1951870173215866, | |
| "learning_rate": 1.567099567099567e-05, | |
| "loss": 0.0826, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 6.0546875, | |
| "grad_norm": 27.74083137512207, | |
| "learning_rate": 1.5454545454545454e-05, | |
| "loss": 0.0697, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "grad_norm": 1.9396530389785767, | |
| "learning_rate": 1.523809523809524e-05, | |
| "loss": 0.0368, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "eval_accuracy": 0.939453125, | |
| "eval_loss": 0.24231916666030884, | |
| "eval_runtime": 5.9487, | |
| "eval_samples_per_second": 172.139, | |
| "eval_steps_per_second": 21.517, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.4453125, | |
| "grad_norm": 0.1338203400373459, | |
| "learning_rate": 1.5021645021645024e-05, | |
| "loss": 0.0482, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 6.640625, | |
| "grad_norm": 0.16084209084510803, | |
| "learning_rate": 1.4805194805194807e-05, | |
| "loss": 0.0382, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 6.8359375, | |
| "grad_norm": 13.326003074645996, | |
| "learning_rate": 1.458874458874459e-05, | |
| "loss": 0.1169, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 7.03125, | |
| "grad_norm": 0.1118919774889946, | |
| "learning_rate": 1.4372294372294374e-05, | |
| "loss": 0.063, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.03125, | |
| "eval_accuracy": 0.927734375, | |
| "eval_loss": 0.30914944410324097, | |
| "eval_runtime": 5.5136, | |
| "eval_samples_per_second": 185.722, | |
| "eval_steps_per_second": 23.215, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.2265625, | |
| "grad_norm": 0.11637281626462936, | |
| "learning_rate": 1.4155844155844157e-05, | |
| "loss": 0.0385, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 7.421875, | |
| "grad_norm": 0.09042005985975266, | |
| "learning_rate": 1.3939393939393942e-05, | |
| "loss": 0.0277, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 7.6171875, | |
| "grad_norm": 0.07495836168527603, | |
| "learning_rate": 1.3722943722943725e-05, | |
| "loss": 0.0214, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 7.8125, | |
| "grad_norm": 0.06756331771612167, | |
| "learning_rate": 1.3506493506493508e-05, | |
| "loss": 0.047, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 7.8125, | |
| "eval_accuracy": 0.90234375, | |
| "eval_loss": 0.3906704783439636, | |
| "eval_runtime": 6.0843, | |
| "eval_samples_per_second": 168.303, | |
| "eval_steps_per_second": 21.038, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.0078125, | |
| "grad_norm": 0.3716655671596527, | |
| "learning_rate": 1.3290043290043291e-05, | |
| "loss": 0.0262, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 8.203125, | |
| "grad_norm": 0.40872329473495483, | |
| "learning_rate": 1.3073593073593074e-05, | |
| "loss": 0.0139, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 8.3984375, | |
| "grad_norm": 0.05600840970873833, | |
| "learning_rate": 1.2861471861471862e-05, | |
| "loss": 0.0161, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 8.59375, | |
| "grad_norm": 0.055402472615242004, | |
| "learning_rate": 1.2645021645021646e-05, | |
| "loss": 0.0127, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 8.59375, | |
| "eval_accuracy": 0.9560546875, | |
| "eval_loss": 0.20022055506706238, | |
| "eval_runtime": 5.6277, | |
| "eval_samples_per_second": 181.956, | |
| "eval_steps_per_second": 22.744, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 8.7890625, | |
| "grad_norm": 0.054736897349357605, | |
| "learning_rate": 1.242857142857143e-05, | |
| "loss": 0.0135, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 8.984375, | |
| "grad_norm": 0.04909529909491539, | |
| "learning_rate": 1.2212121212121213e-05, | |
| "loss": 0.0209, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 9.1796875, | |
| "grad_norm": 0.04906412959098816, | |
| "learning_rate": 1.1995670995670996e-05, | |
| "loss": 0.0107, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 9.375, | |
| "grad_norm": 0.04470803588628769, | |
| "learning_rate": 1.1779220779220779e-05, | |
| "loss": 0.0102, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.375, | |
| "eval_accuracy": 0.9306640625, | |
| "eval_loss": 0.30005937814712524, | |
| "eval_runtime": 6.136, | |
| "eval_samples_per_second": 166.885, | |
| "eval_steps_per_second": 20.861, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.5703125, | |
| "grad_norm": 0.04809673875570297, | |
| "learning_rate": 1.1562770562770564e-05, | |
| "loss": 0.01, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 9.765625, | |
| "grad_norm": 0.053658001124858856, | |
| "learning_rate": 1.1346320346320347e-05, | |
| "loss": 0.0096, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 9.9609375, | |
| "grad_norm": 0.05094292759895325, | |
| "learning_rate": 1.112987012987013e-05, | |
| "loss": 0.0085, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 10.15625, | |
| "grad_norm": 0.03785080462694168, | |
| "learning_rate": 1.0913419913419913e-05, | |
| "loss": 0.0086, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 10.15625, | |
| "eval_accuracy": 0.951171875, | |
| "eval_loss": 0.19980870187282562, | |
| "eval_runtime": 5.5065, | |
| "eval_samples_per_second": 185.964, | |
| "eval_steps_per_second": 23.245, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 10.3515625, | |
| "grad_norm": 0.048504624515771866, | |
| "learning_rate": 1.0696969696969696e-05, | |
| "loss": 0.0078, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 10.546875, | |
| "grad_norm": 0.040160298347473145, | |
| "learning_rate": 1.0480519480519481e-05, | |
| "loss": 0.0136, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 10.7421875, | |
| "grad_norm": 0.03454400226473808, | |
| "learning_rate": 1.0264069264069264e-05, | |
| "loss": 0.0077, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 10.9375, | |
| "grad_norm": 0.03354168310761452, | |
| "learning_rate": 1.0047619047619048e-05, | |
| "loss": 0.0073, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 10.9375, | |
| "eval_accuracy": 0.958984375, | |
| "eval_loss": 0.19315651059150696, | |
| "eval_runtime": 6.0024, | |
| "eval_samples_per_second": 170.598, | |
| "eval_steps_per_second": 21.325, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 11.1328125, | |
| "grad_norm": 0.033880043774843216, | |
| "learning_rate": 9.831168831168832e-06, | |
| "loss": 0.0071, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 11.328125, | |
| "grad_norm": 0.032037124037742615, | |
| "learning_rate": 9.614718614718616e-06, | |
| "loss": 0.0066, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 11.5234375, | |
| "grad_norm": 0.0329328328371048, | |
| "learning_rate": 9.398268398268399e-06, | |
| "loss": 0.0066, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 11.71875, | |
| "grad_norm": 0.02779380977153778, | |
| "learning_rate": 9.181818181818184e-06, | |
| "loss": 0.0064, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 11.71875, | |
| "eval_accuracy": 0.9560546875, | |
| "eval_loss": 0.19875724613666534, | |
| "eval_runtime": 5.8842, | |
| "eval_samples_per_second": 174.025, | |
| "eval_steps_per_second": 21.753, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 11.9140625, | |
| "grad_norm": 0.028905285522341728, | |
| "learning_rate": 8.965367965367967e-06, | |
| "loss": 0.0063, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 12.109375, | |
| "grad_norm": 0.027350405231118202, | |
| "learning_rate": 8.74891774891775e-06, | |
| "loss": 0.0058, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 12.3046875, | |
| "grad_norm": 0.026442719623446465, | |
| "learning_rate": 8.532467532467533e-06, | |
| "loss": 0.0059, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "grad_norm": 0.028089182451367378, | |
| "learning_rate": 8.316017316017316e-06, | |
| "loss": 0.0056, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "eval_accuracy": 0.9580078125, | |
| "eval_loss": 0.19929003715515137, | |
| "eval_runtime": 6.1143, | |
| "eval_samples_per_second": 167.475, | |
| "eval_steps_per_second": 20.934, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 12.6953125, | |
| "grad_norm": 0.02532665804028511, | |
| "learning_rate": 8.099567099567101e-06, | |
| "loss": 0.0053, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 12.890625, | |
| "grad_norm": 0.024192778393626213, | |
| "learning_rate": 7.883116883116884e-06, | |
| "loss": 0.0055, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 13.0859375, | |
| "grad_norm": 0.028951533138751984, | |
| "learning_rate": 7.666666666666667e-06, | |
| "loss": 0.0053, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 13.28125, | |
| "grad_norm": 0.024818725883960724, | |
| "learning_rate": 7.450216450216451e-06, | |
| "loss": 0.0049, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 13.28125, | |
| "eval_accuracy": 0.958984375, | |
| "eval_loss": 0.2046782374382019, | |
| "eval_runtime": 5.5223, | |
| "eval_samples_per_second": 185.43, | |
| "eval_steps_per_second": 23.179, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 13.28125, | |
| "step": 3400, | |
| "total_flos": 4.2156398005714944e+18, | |
| "train_loss": 0.23371873112285838, | |
| "train_runtime": 1043.0698, | |
| "train_samples_per_second": 78.537, | |
| "train_steps_per_second": 4.909 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 5120, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 200, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.2156398005714944e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |