| { | |
| "best_metric": 0.11504850536584854, | |
| "best_model_checkpoint": "./vit-base-obipix_birdID-v1/checkpoint-28000", | |
| "epoch": 5.0, | |
| "eval_steps": 1000, | |
| "global_step": 28430, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 109984.9296875, | |
| "learning_rate": 0.00019296517762926486, | |
| "loss": 6.9257, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "eval_accuracy": 0.16375981063909306, | |
| "eval_loss": 5.382956027984619, | |
| "eval_runtime": 503.1163, | |
| "eval_samples_per_second": 127.636, | |
| "eval_steps_per_second": 7.978, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 176902.25, | |
| "learning_rate": 0.00018593035525852974, | |
| "loss": 3.9727, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "eval_accuracy": 0.4797246792076741, | |
| "eval_loss": 2.769526243209839, | |
| "eval_runtime": 503.5958, | |
| "eval_samples_per_second": 127.515, | |
| "eval_steps_per_second": 7.971, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 220305.0625, | |
| "learning_rate": 0.0001788955328877946, | |
| "loss": 2.057, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.693612183879407, | |
| "eval_loss": 1.5069622993469238, | |
| "eval_runtime": 494.5536, | |
| "eval_samples_per_second": 129.846, | |
| "eval_steps_per_second": 8.116, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 227490.640625, | |
| "learning_rate": 0.00017186071051705947, | |
| "loss": 1.2103, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_accuracy": 0.7841815123956646, | |
| "eval_loss": 0.9726743102073669, | |
| "eval_runtime": 494.0168, | |
| "eval_samples_per_second": 129.987, | |
| "eval_steps_per_second": 8.125, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 246928.109375, | |
| "learning_rate": 0.00016482588814632432, | |
| "loss": 0.8513, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_accuracy": 0.8318487604335368, | |
| "eval_loss": 0.710105836391449, | |
| "eval_runtime": 492.9112, | |
| "eval_samples_per_second": 130.279, | |
| "eval_steps_per_second": 8.143, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "grad_norm": 182599.234375, | |
| "learning_rate": 0.00015779106577558917, | |
| "loss": 0.5836, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "eval_accuracy": 0.8561106266351065, | |
| "eval_loss": 0.57969069480896, | |
| "eval_runtime": 497.4445, | |
| "eval_samples_per_second": 129.092, | |
| "eval_steps_per_second": 8.069, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 174586.96875, | |
| "learning_rate": 0.00015075624340485404, | |
| "loss": 0.3545, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "eval_accuracy": 0.8729755824093683, | |
| "eval_loss": 0.5065863132476807, | |
| "eval_runtime": 490.3364, | |
| "eval_samples_per_second": 130.963, | |
| "eval_steps_per_second": 8.186, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 164026.03125, | |
| "learning_rate": 0.0001437214210341189, | |
| "loss": 0.314, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_accuracy": 0.8818363024791329, | |
| "eval_loss": 0.4521096646785736, | |
| "eval_runtime": 487.972, | |
| "eval_samples_per_second": 131.598, | |
| "eval_steps_per_second": 8.226, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 164116.1875, | |
| "learning_rate": 0.00013668659866338377, | |
| "loss": 0.2858, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_accuracy": 0.8959760807275445, | |
| "eval_loss": 0.39154955744743347, | |
| "eval_runtime": 498.2235, | |
| "eval_samples_per_second": 128.89, | |
| "eval_steps_per_second": 8.057, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 142166.859375, | |
| "learning_rate": 0.00012965177629264862, | |
| "loss": 0.2482, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "eval_accuracy": 0.9055687056185374, | |
| "eval_loss": 0.3563849925994873, | |
| "eval_runtime": 491.8161, | |
| "eval_samples_per_second": 130.569, | |
| "eval_steps_per_second": 8.162, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "grad_norm": 160608.578125, | |
| "learning_rate": 0.00012261695392191347, | |
| "loss": 0.2192, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "eval_accuracy": 0.9148343092064283, | |
| "eval_loss": 0.3131472170352936, | |
| "eval_runtime": 494.6571, | |
| "eval_samples_per_second": 129.819, | |
| "eval_steps_per_second": 8.115, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "grad_norm": 72832.3359375, | |
| "learning_rate": 0.00011558213155117834, | |
| "loss": 0.1271, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "eval_accuracy": 0.9206739753332502, | |
| "eval_loss": 0.29156896471977234, | |
| "eval_runtime": 490.4403, | |
| "eval_samples_per_second": 130.935, | |
| "eval_steps_per_second": 8.184, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 63710.625, | |
| "learning_rate": 0.0001085473091804432, | |
| "loss": 0.0779, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "eval_accuracy": 0.9259997508409119, | |
| "eval_loss": 0.272681325674057, | |
| "eval_runtime": 493.6688, | |
| "eval_samples_per_second": 130.079, | |
| "eval_steps_per_second": 8.131, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "grad_norm": 108940.078125, | |
| "learning_rate": 0.00010151248680970806, | |
| "loss": 0.0749, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "eval_accuracy": 0.9308894979444375, | |
| "eval_loss": 0.259682297706604, | |
| "eval_runtime": 489.3469, | |
| "eval_samples_per_second": 131.228, | |
| "eval_steps_per_second": 8.203, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "grad_norm": 118623.28125, | |
| "learning_rate": 9.447766443897293e-05, | |
| "loss": 0.0682, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "eval_accuracy": 0.9354677961878659, | |
| "eval_loss": 0.2414509356021881, | |
| "eval_runtime": 498.6017, | |
| "eval_samples_per_second": 128.792, | |
| "eval_steps_per_second": 8.051, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 59107.12890625, | |
| "learning_rate": 8.744284206823779e-05, | |
| "loss": 0.0615, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "eval_accuracy": 0.9384577052447988, | |
| "eval_loss": 0.226842001080513, | |
| "eval_runtime": 489.9279, | |
| "eval_samples_per_second": 131.072, | |
| "eval_steps_per_second": 8.193, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 104112.5390625, | |
| "learning_rate": 8.040801969750264e-05, | |
| "loss": 0.0566, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_accuracy": 0.9439703500685187, | |
| "eval_loss": 0.2083783596754074, | |
| "eval_runtime": 498.3818, | |
| "eval_samples_per_second": 128.849, | |
| "eval_steps_per_second": 8.054, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "grad_norm": 65992.59375, | |
| "learning_rate": 7.33731973267675e-05, | |
| "loss": 0.0197, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "eval_accuracy": 0.9474897221876168, | |
| "eval_loss": 0.19510340690612793, | |
| "eval_runtime": 491.467, | |
| "eval_samples_per_second": 130.662, | |
| "eval_steps_per_second": 8.167, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.34, | |
| "grad_norm": 27417.0, | |
| "learning_rate": 6.633837495603237e-05, | |
| "loss": 0.0158, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.34, | |
| "eval_accuracy": 0.9513049707238072, | |
| "eval_loss": 0.1843044012784958, | |
| "eval_runtime": 491.4448, | |
| "eval_samples_per_second": 130.668, | |
| "eval_steps_per_second": 8.168, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.52, | |
| "grad_norm": 9745.1337890625, | |
| "learning_rate": 5.930355258529723e-05, | |
| "loss": 0.0145, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.52, | |
| "eval_accuracy": 0.9541391553506914, | |
| "eval_loss": 0.17461912333965302, | |
| "eval_runtime": 497.3213, | |
| "eval_samples_per_second": 129.124, | |
| "eval_steps_per_second": 8.071, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "grad_norm": 12192.4462890625, | |
| "learning_rate": 5.226873021456209e-05, | |
| "loss": 0.0118, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "eval_accuracy": 0.957269216394668, | |
| "eval_loss": 0.1649465709924698, | |
| "eval_runtime": 491.5529, | |
| "eval_samples_per_second": 130.639, | |
| "eval_steps_per_second": 8.166, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 18333.3828125, | |
| "learning_rate": 4.523390784382695e-05, | |
| "loss": 0.0103, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "eval_accuracy": 0.9599009592624891, | |
| "eval_loss": 0.1530667096376419, | |
| "eval_runtime": 499.5499, | |
| "eval_samples_per_second": 128.548, | |
| "eval_steps_per_second": 8.035, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 4.05, | |
| "grad_norm": 1477.312744140625, | |
| "learning_rate": 3.8199085473091805e-05, | |
| "loss": 0.006, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.05, | |
| "eval_accuracy": 0.9643546779618787, | |
| "eval_loss": 0.13792124390602112, | |
| "eval_runtime": 499.2627, | |
| "eval_samples_per_second": 128.622, | |
| "eval_steps_per_second": 8.04, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.22, | |
| "grad_norm": 903.4931030273438, | |
| "learning_rate": 3.116426310235667e-05, | |
| "loss": 0.0016, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.22, | |
| "eval_accuracy": 0.9667528341846269, | |
| "eval_loss": 0.13162368535995483, | |
| "eval_runtime": 486.4405, | |
| "eval_samples_per_second": 132.012, | |
| "eval_steps_per_second": 8.252, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "grad_norm": 1672.1170654296875, | |
| "learning_rate": 2.412944073162153e-05, | |
| "loss": 0.0013, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "eval_accuracy": 0.9685903824592001, | |
| "eval_loss": 0.12647932767868042, | |
| "eval_runtime": 488.945, | |
| "eval_samples_per_second": 131.336, | |
| "eval_steps_per_second": 8.21, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.57, | |
| "grad_norm": 179.9572296142578, | |
| "learning_rate": 1.709461836088639e-05, | |
| "loss": 0.0014, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.57, | |
| "eval_accuracy": 0.9697427432415597, | |
| "eval_loss": 0.1232375055551529, | |
| "eval_runtime": 488.3862, | |
| "eval_samples_per_second": 131.486, | |
| "eval_steps_per_second": 8.219, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "grad_norm": 538.2374877929688, | |
| "learning_rate": 1.005979599015125e-05, | |
| "loss": 0.0009, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "eval_accuracy": 0.9711598355550018, | |
| "eval_loss": 0.11887777596712112, | |
| "eval_runtime": 490.273, | |
| "eval_samples_per_second": 130.98, | |
| "eval_steps_per_second": 8.187, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.92, | |
| "grad_norm": 845.1385498046875, | |
| "learning_rate": 3.02497361941611e-06, | |
| "loss": 0.001, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 4.92, | |
| "eval_accuracy": 0.9719696025912545, | |
| "eval_loss": 0.11504850536584854, | |
| "eval_runtime": 495.362, | |
| "eval_samples_per_second": 129.634, | |
| "eval_steps_per_second": 8.103, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 28430, | |
| "total_flos": 1.4469915597827883e+20, | |
| "train_loss": 0.618133510268158, | |
| "train_runtime": 34089.5693, | |
| "train_samples_per_second": 53.373, | |
| "train_steps_per_second": 0.834 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9719696025912545, | |
| "eval_loss": 0.11504850536584854, | |
| "eval_runtime": 489.2058, | |
| "eval_samples_per_second": 131.266, | |
| "eval_steps_per_second": 8.205, | |
| "step": 28430 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 28430, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 1000, | |
| "total_flos": 1.4469915597827883e+20, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |