| { | |
| "best_metric": 0.8391695022583008, | |
| "best_model_checkpoint": "./quickdraw-ConvNeXT-Tiny-Finetune/checkpoint-25000", | |
| "epoch": 7.0, | |
| "eval_steps": 5000, | |
| "global_step": 61530, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11376564277588168, | |
| "grad_norm": 7.832313537597656, | |
| "learning_rate": 0.00015952, | |
| "loss": 3.8212, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.22753128555176336, | |
| "grad_norm": 2.515751600265503, | |
| "learning_rate": 0.00031952, | |
| "loss": 1.4714, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.3412969283276451, | |
| "grad_norm": 1.605010747909546, | |
| "learning_rate": 0.00047952000000000004, | |
| "loss": 1.2045, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.4550625711035267, | |
| "grad_norm": 1.3963702917099, | |
| "learning_rate": 0.0006395200000000001, | |
| "loss": 1.1247, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5688282138794084, | |
| "grad_norm": 1.0773191452026367, | |
| "learning_rate": 0.00079952, | |
| "loss": 1.0895, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.5688282138794084, | |
| "eval_accuracy": 0.727772, | |
| "eval_loss": 1.0618326663970947, | |
| "eval_runtime": 25.193, | |
| "eval_samples_per_second": 9923.407, | |
| "eval_steps_per_second": 19.41, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.6825938566552902, | |
| "grad_norm": 0.9033334255218506, | |
| "learning_rate": 0.0007993861659457679, | |
| "loss": 1.0466, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.7963594994311718, | |
| "grad_norm": 0.9264630079269409, | |
| "learning_rate": 0.000797541628656647, | |
| "loss": 1.0055, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.9101251422070534, | |
| "grad_norm": 0.8748724460601807, | |
| "learning_rate": 0.0007944683860683685, | |
| "loss": 0.9762, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.023890784982935, | |
| "grad_norm": 0.7623477578163147, | |
| "learning_rate": 0.0007901820532646013, | |
| "loss": 0.9451, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.1376564277588168, | |
| "grad_norm": 0.7780616283416748, | |
| "learning_rate": 0.000784687285963723, | |
| "loss": 0.8912, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1376564277588168, | |
| "eval_accuracy": 0.760228, | |
| "eval_loss": 0.9315244555473328, | |
| "eval_runtime": 25.8226, | |
| "eval_samples_per_second": 9681.46, | |
| "eval_steps_per_second": 18.937, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.2514220705346986, | |
| "grad_norm": 0.8434277772903442, | |
| "learning_rate": 0.0007780120029676767, | |
| "loss": 0.8896, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.36518771331058, | |
| "grad_norm": 0.7569168210029602, | |
| "learning_rate": 0.0007701634548188583, | |
| "loss": 0.8879, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.4789533560864618, | |
| "grad_norm": 0.8010281920433044, | |
| "learning_rate": 0.000761171966395704, | |
| "loss": 0.8761, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.5927189988623436, | |
| "grad_norm": 0.6517601013183594, | |
| "learning_rate": 0.0007510759539838813, | |
| "loss": 0.8712, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.7064846416382253, | |
| "grad_norm": 1.0584135055541992, | |
| "learning_rate": 0.0007398863836452142, | |
| "loss": 0.8658, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.7064846416382253, | |
| "eval_accuracy": 0.770068, | |
| "eval_loss": 0.887133777141571, | |
| "eval_runtime": 25.3772, | |
| "eval_samples_per_second": 9851.358, | |
| "eval_steps_per_second": 19.269, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.820250284414107, | |
| "grad_norm": 1.1719928979873657, | |
| "learning_rate": 0.0007276473584467834, | |
| "loss": 0.863, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.9340159271899886, | |
| "grad_norm": 0.9330978393554688, | |
| "learning_rate": 0.0007143966683828085, | |
| "loss": 0.8548, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.04778156996587, | |
| "grad_norm": 0.6235845685005188, | |
| "learning_rate": 0.0007001899190878672, | |
| "loss": 0.8088, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.161547212741752, | |
| "grad_norm": 0.6521238684654236, | |
| "learning_rate": 0.0006850425416138172, | |
| "loss": 0.7471, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.2753128555176336, | |
| "grad_norm": 0.6797402501106262, | |
| "learning_rate": 0.0006690314994804036, | |
| "loss": 0.7558, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.2753128555176336, | |
| "eval_accuracy": 0.77796, | |
| "eval_loss": 0.8652966022491455, | |
| "eval_runtime": 25.3581, | |
| "eval_samples_per_second": 9858.786, | |
| "eval_steps_per_second": 19.284, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.3890784982935154, | |
| "grad_norm": 0.9936743974685669, | |
| "learning_rate": 0.0006521741835218227, | |
| "loss": 0.7588, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.502844141069397, | |
| "grad_norm": 1.0131540298461914, | |
| "learning_rate": 0.000634556245420806, | |
| "loss": 0.7554, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.616609783845279, | |
| "grad_norm": 0.7528037428855896, | |
| "learning_rate": 0.0006161968213858429, | |
| "loss": 0.7605, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 2.73037542662116, | |
| "grad_norm": 0.6982817649841309, | |
| "learning_rate": 0.0005971891952805355, | |
| "loss": 0.7584, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.8441410693970424, | |
| "grad_norm": 0.6886234283447266, | |
| "learning_rate": 0.000577554012761257, | |
| "loss": 0.7578, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.8441410693970424, | |
| "eval_accuracy": 0.782604, | |
| "eval_loss": 0.8391695022583008, | |
| "eval_runtime": 25.446, | |
| "eval_samples_per_second": 9824.738, | |
| "eval_steps_per_second": 19.217, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.9579067121729237, | |
| "grad_norm": 0.7857435345649719, | |
| "learning_rate": 0.0005573910397947634, | |
| "loss": 0.7511, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 3.0716723549488054, | |
| "grad_norm": 0.7772754430770874, | |
| "learning_rate": 0.0005367430674286641, | |
| "loss": 0.664, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 3.185437997724687, | |
| "grad_norm": 0.8711711168289185, | |
| "learning_rate": 0.0005156524423003168, | |
| "loss": 0.6158, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 3.299203640500569, | |
| "grad_norm": 0.7758294343948364, | |
| "learning_rate": 0.0004942047213203677, | |
| "loss": 0.621, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 3.4129692832764507, | |
| "grad_norm": 0.7045032978057861, | |
| "learning_rate": 0.00047246612783901975, | |
| "loss": 0.6249, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 3.4129692832764507, | |
| "eval_accuracy": 0.784164, | |
| "eval_loss": 0.8584216237068176, | |
| "eval_runtime": 25.4184, | |
| "eval_samples_per_second": 9835.409, | |
| "eval_steps_per_second": 19.238, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 3.526734926052332, | |
| "grad_norm": 0.7496004104614258, | |
| "learning_rate": 0.00045052583490800234, | |
| "loss": 0.6296, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 3.640500568828214, | |
| "grad_norm": 1.1859596967697144, | |
| "learning_rate": 0.00042840767358049813, | |
| "loss": 0.6323, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 3.7542662116040955, | |
| "grad_norm": 0.6882604360580444, | |
| "learning_rate": 0.0004062240257885585, | |
| "loss": 0.6255, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 3.868031854379977, | |
| "grad_norm": 0.7085530757904053, | |
| "learning_rate": 0.00038399898691194754, | |
| "loss": 0.6267, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 3.981797497155859, | |
| "grad_norm": 0.7478043437004089, | |
| "learning_rate": 0.00036184548192539, | |
| "loss": 0.6243, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 3.981797497155859, | |
| "eval_accuracy": 0.7882, | |
| "eval_loss": 0.8410258293151855, | |
| "eval_runtime": 25.3906, | |
| "eval_samples_per_second": 9846.165, | |
| "eval_steps_per_second": 19.259, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 4.09556313993174, | |
| "grad_norm": 0.9012534022331238, | |
| "learning_rate": 0.0003397875734682125, | |
| "loss": 0.481, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 4.2093287827076225, | |
| "grad_norm": 0.8275543451309204, | |
| "learning_rate": 0.0003179373373288462, | |
| "loss": 0.4521, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 4.323094425483504, | |
| "grad_norm": 0.7898097038269043, | |
| "learning_rate": 0.0002963185067428097, | |
| "loss": 0.4555, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 4.436860068259386, | |
| "grad_norm": 0.853428840637207, | |
| "learning_rate": 0.00027504092655271806, | |
| "loss": 0.4603, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 4.550625711035267, | |
| "grad_norm": 0.857472836971283, | |
| "learning_rate": 0.0002541277079880658, | |
| "loss": 0.4585, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 4.550625711035267, | |
| "eval_accuracy": 0.782792, | |
| "eval_loss": 0.9402379989624023, | |
| "eval_runtime": 25.4609, | |
| "eval_samples_per_second": 9818.981, | |
| "eval_steps_per_second": 19.206, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 4.664391353811149, | |
| "grad_norm": 0.7866658568382263, | |
| "learning_rate": 0.00023368511069100407, | |
| "loss": 0.4596, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 4.778156996587031, | |
| "grad_norm": 0.8958695530891418, | |
| "learning_rate": 0.0002137353389513713, | |
| "loss": 0.4585, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 4.891922639362912, | |
| "grad_norm": 0.8633857369422913, | |
| "learning_rate": 0.0001943797571574129, | |
| "loss": 0.4587, | |
| "step": 43000 | |
| }, | |
| { | |
| "epoch": 5.005688282138794, | |
| "grad_norm": 0.6427882313728333, | |
| "learning_rate": 0.0001756393889071532, | |
| "loss": 0.4445, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 5.1194539249146755, | |
| "grad_norm": 0.8369711637496948, | |
| "learning_rate": 0.00015760945363407598, | |
| "loss": 0.2934, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 5.1194539249146755, | |
| "eval_accuracy": 0.775848, | |
| "eval_loss": 1.0586884021759033, | |
| "eval_runtime": 25.6083, | |
| "eval_samples_per_second": 9762.449, | |
| "eval_steps_per_second": 19.095, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 5.233219567690558, | |
| "grad_norm": 0.8710008859634399, | |
| "learning_rate": 0.00014030953504874178, | |
| "loss": 0.2876, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 5.346985210466439, | |
| "grad_norm": 1.0742043256759644, | |
| "learning_rate": 0.00012381145329356995, | |
| "loss": 0.286, | |
| "step": 47000 | |
| }, | |
| { | |
| "epoch": 5.460750853242321, | |
| "grad_norm": 1.2184427976608276, | |
| "learning_rate": 0.00010818135193410835, | |
| "loss": 0.2893, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 5.5745164960182025, | |
| "grad_norm": 1.0682432651519775, | |
| "learning_rate": 9.343620803524462e-05, | |
| "loss": 0.2883, | |
| "step": 49000 | |
| }, | |
| { | |
| "epoch": 5.688282138794084, | |
| "grad_norm": 0.9291980862617493, | |
| "learning_rate": 7.965094137603242e-05, | |
| "loss": 0.2869, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 5.688282138794084, | |
| "eval_accuracy": 0.774452, | |
| "eval_loss": 1.1279938220977783, | |
| "eval_runtime": 25.7253, | |
| "eval_samples_per_second": 9718.054, | |
| "eval_steps_per_second": 19.009, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 5.802047781569966, | |
| "grad_norm": 0.9417611360549927, | |
| "learning_rate": 6.684052520336171e-05, | |
| "loss": 0.2845, | |
| "step": 51000 | |
| }, | |
| { | |
| "epoch": 5.915813424345847, | |
| "grad_norm": 0.8925788998603821, | |
| "learning_rate": 5.507004898369141e-05, | |
| "loss": 0.2845, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 6.0295790671217295, | |
| "grad_norm": 0.6515630483627319, | |
| "learning_rate": 4.435229754380843e-05, | |
| "loss": 0.2576, | |
| "step": 53000 | |
| }, | |
| { | |
| "epoch": 6.143344709897611, | |
| "grad_norm": 0.6938067674636841, | |
| "learning_rate": 3.474172756282932e-05, | |
| "loss": 0.187, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 6.257110352673493, | |
| "grad_norm": 0.7021601796150208, | |
| "learning_rate": 2.6248777825917638e-05, | |
| "loss": 0.1837, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 6.257110352673493, | |
| "eval_accuracy": 0.768376, | |
| "eval_loss": 1.2402174472808838, | |
| "eval_runtime": 25.1613, | |
| "eval_samples_per_second": 9935.898, | |
| "eval_steps_per_second": 19.435, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 6.370875995449374, | |
| "grad_norm": 1.0282700061798096, | |
| "learning_rate": 1.891660083956399e-05, | |
| "loss": 0.1852, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 6.484641638225256, | |
| "grad_norm": 0.7588028311729431, | |
| "learning_rate": 1.2753160649656704e-05, | |
| "loss": 0.1838, | |
| "step": 57000 | |
| }, | |
| { | |
| "epoch": 6.598407281001138, | |
| "grad_norm": 0.9116990566253662, | |
| "learning_rate": 7.789773571614944e-06, | |
| "loss": 0.1857, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 6.712172923777019, | |
| "grad_norm": 0.8666180372238159, | |
| "learning_rate": 4.031830724955432e-06, | |
| "loss": 0.1855, | |
| "step": 59000 | |
| }, | |
| { | |
| "epoch": 6.825938566552901, | |
| "grad_norm": 0.7389401793479919, | |
| "learning_rate": 1.4984261415641243e-06, | |
| "loss": 0.1858, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 6.825938566552901, | |
| "eval_accuracy": 0.767868, | |
| "eval_loss": 1.2570157051086426, | |
| "eval_runtime": 25.8837, | |
| "eval_samples_per_second": 9658.593, | |
| "eval_steps_per_second": 18.892, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 6.939704209328783, | |
| "grad_norm": 0.871012806892395, | |
| "learning_rate": 1.9231154857375366e-07, | |
| "loss": 0.1816, | |
| "step": 61000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 61530, | |
| "total_flos": 4.160918550096e+18, | |
| "train_loss": 0.6576268009338925, | |
| "train_runtime": 10956.031, | |
| "train_samples_per_second": 2875.129, | |
| "train_steps_per_second": 5.616 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 61530, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 7, | |
| "save_steps": 5000, | |
| "total_flos": 4.160918550096e+18, | |
| "train_batch_size": 512, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |