| { | |
| "best_metric": 0.23698414862155914, | |
| "best_model_checkpoint": "./convnext-tiny-new-1e-4-batch-32/checkpoint-5500", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 5500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 18.67238426208496, | |
| "learning_rate": 9.991845519630678e-05, | |
| "loss": 2.1347, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 24.724580764770508, | |
| "learning_rate": 9.967408676742751e-05, | |
| "loss": 1.1347, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 23.250165939331055, | |
| "learning_rate": 9.926769179238466e-05, | |
| "loss": 0.9363, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 20.889720916748047, | |
| "learning_rate": 9.870059584711668e-05, | |
| "loss": 0.8421, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 16.967330932617188, | |
| "learning_rate": 9.797464868072488e-05, | |
| "loss": 0.7731, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.88389662027833, | |
| "eval_loss": 0.42373502254486084, | |
| "eval_runtime": 76.9715, | |
| "eval_samples_per_second": 32.674, | |
| "eval_steps_per_second": 1.026, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "grad_norm": 19.625564575195312, | |
| "learning_rate": 9.709221818197624e-05, | |
| "loss": 0.7413, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "grad_norm": 16.88126564025879, | |
| "learning_rate": 9.60561826557425e-05, | |
| "loss": 0.6635, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 14.250543594360352, | |
| "learning_rate": 9.486992143456792e-05, | |
| "loss": 0.6125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 14.460857391357422, | |
| "learning_rate": 9.353730385598887e-05, | |
| "loss": 0.6703, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 16.239057540893555, | |
| "learning_rate": 9.206267664155907e-05, | |
| "loss": 0.5786, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 25.278282165527344, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.6206, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8930417495029821, | |
| "eval_loss": 0.37559983134269714, | |
| "eval_runtime": 77.5566, | |
| "eval_samples_per_second": 32.428, | |
| "eval_steps_per_second": 1.019, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 8.611919403076172, | |
| "learning_rate": 8.870708053195413e-05, | |
| "loss": 0.5204, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "grad_norm": 13.736300468444824, | |
| "learning_rate": 8.683705689382024e-05, | |
| "loss": 0.5279, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 12.42533016204834, | |
| "learning_rate": 8.484687843276469e-05, | |
| "loss": 0.489, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "grad_norm": 15.548003196716309, | |
| "learning_rate": 8.274303669726426e-05, | |
| "loss": 0.5077, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "grad_norm": 14.506275177001953, | |
| "learning_rate": 8.053239398177191e-05, | |
| "loss": 0.5014, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9073558648111332, | |
| "eval_loss": 0.3272249698638916, | |
| "eval_runtime": 75.9338, | |
| "eval_samples_per_second": 33.121, | |
| "eval_steps_per_second": 1.04, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "grad_norm": 18.079133987426758, | |
| "learning_rate": 7.822216094333847e-05, | |
| "loss": 0.4807, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "grad_norm": 13.355294227600098, | |
| "learning_rate": 7.58198730819481e-05, | |
| "loss": 0.4562, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "grad_norm": 13.22020149230957, | |
| "learning_rate": 7.333336616128369e-05, | |
| "loss": 0.4124, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "grad_norm": 14.023577690124512, | |
| "learning_rate": 7.077075065009433e-05, | |
| "loss": 0.4643, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.82, | |
| "grad_norm": 13.74384880065918, | |
| "learning_rate": 6.814038526753205e-05, | |
| "loss": 0.444, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 16.15762710571289, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.3892, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9280318091451292, | |
| "eval_loss": 0.27286219596862793, | |
| "eval_runtime": 76.623, | |
| "eval_samples_per_second": 32.823, | |
| "eval_steps_per_second": 1.031, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "grad_norm": 16.897714614868164, | |
| "learning_rate": 6.271091670967436e-05, | |
| "loss": 0.3776, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "grad_norm": 9.421957015991211, | |
| "learning_rate": 5.992952333228728e-05, | |
| "loss": 0.3594, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.55, | |
| "grad_norm": 9.755322456359863, | |
| "learning_rate": 5.7115741913664264e-05, | |
| "loss": 0.3712, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "grad_norm": 10.371528625488281, | |
| "learning_rate": 5.427875042394199e-05, | |
| "loss": 0.3862, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 4.91, | |
| "grad_norm": 8.203412055969238, | |
| "learning_rate": 5.142780253968481e-05, | |
| "loss": 0.3819, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9204771371769384, | |
| "eval_loss": 0.295335978269577, | |
| "eval_runtime": 78.1287, | |
| "eval_samples_per_second": 32.19, | |
| "eval_steps_per_second": 1.011, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 5.09, | |
| "grad_norm": 7.4753098487854, | |
| "learning_rate": 4.85721974603152e-05, | |
| "loss": 0.3352, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 5.27, | |
| "grad_norm": 9.154393196105957, | |
| "learning_rate": 4.5721249576058027e-05, | |
| "loss": 0.3283, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "grad_norm": 11.729723930358887, | |
| "learning_rate": 4.288425808633575e-05, | |
| "loss": 0.3199, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 5.64, | |
| "grad_norm": 11.630327224731445, | |
| "learning_rate": 4.007047666771274e-05, | |
| "loss": 0.3143, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "grad_norm": 4.333253860473633, | |
| "learning_rate": 3.728908329032567e-05, | |
| "loss": 0.3231, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 18.835886001586914, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.3315, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9260437375745527, | |
| "eval_loss": 0.2783779799938202, | |
| "eval_runtime": 76.4348, | |
| "eval_samples_per_second": 32.904, | |
| "eval_steps_per_second": 1.034, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 6.18, | |
| "grad_norm": 5.403916835784912, | |
| "learning_rate": 3.1859614732467954e-05, | |
| "loss": 0.2781, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "grad_norm": 11.396268844604492, | |
| "learning_rate": 2.9229249349905684e-05, | |
| "loss": 0.2865, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 6.55, | |
| "grad_norm": 15.219473838806152, | |
| "learning_rate": 2.6666633838716314e-05, | |
| "loss": 0.2839, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "grad_norm": 11.592494010925293, | |
| "learning_rate": 2.418012691805191e-05, | |
| "loss": 0.2549, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 6.91, | |
| "grad_norm": 10.148946762084961, | |
| "learning_rate": 2.1777839056661554e-05, | |
| "loss": 0.281, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.932803180914513, | |
| "eval_loss": 0.2535076141357422, | |
| "eval_runtime": 77.3926, | |
| "eval_samples_per_second": 32.497, | |
| "eval_steps_per_second": 1.021, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 7.09, | |
| "grad_norm": 10.556535720825195, | |
| "learning_rate": 1.946760601822809e-05, | |
| "loss": 0.2247, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 7.27, | |
| "grad_norm": 12.908746719360352, | |
| "learning_rate": 1.725696330273575e-05, | |
| "loss": 0.2302, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 7.45, | |
| "grad_norm": 11.96335506439209, | |
| "learning_rate": 1.5153121567235335e-05, | |
| "loss": 0.2479, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 7.64, | |
| "grad_norm": 13.116978645324707, | |
| "learning_rate": 1.3162943106179749e-05, | |
| "loss": 0.2423, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 7.82, | |
| "grad_norm": 6.12368631362915, | |
| "learning_rate": 1.1292919468045877e-05, | |
| "loss": 0.2317, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 31.24999237060547, | |
| "learning_rate": 9.549150281252633e-06, | |
| "loss": 0.217, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9343936381709742, | |
| "eval_loss": 0.24977222084999084, | |
| "eval_runtime": 75.2887, | |
| "eval_samples_per_second": 33.405, | |
| "eval_steps_per_second": 1.049, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 8.18, | |
| "grad_norm": 12.625907897949219, | |
| "learning_rate": 7.937323358440935e-06, | |
| "loss": 0.208, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 8.36, | |
| "grad_norm": 10.004912376403809, | |
| "learning_rate": 6.462696144011149e-06, | |
| "loss": 0.2105, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 8.55, | |
| "grad_norm": 18.372529983520508, | |
| "learning_rate": 5.13007856543209e-06, | |
| "loss": 0.2038, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 8.73, | |
| "grad_norm": 13.66908073425293, | |
| "learning_rate": 3.9438173442575e-06, | |
| "loss": 0.2216, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 8.91, | |
| "grad_norm": 3.5467135906219482, | |
| "learning_rate": 2.9077818180237693e-06, | |
| "loss": 0.2152, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.9399602385685885, | |
| "eval_loss": 0.24052982032299042, | |
| "eval_runtime": 72.694, | |
| "eval_samples_per_second": 34.597, | |
| "eval_steps_per_second": 1.087, | |
| "step": 4950 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "grad_norm": 7.110690116882324, | |
| "learning_rate": 2.0253513192751373e-06, | |
| "loss": 0.1938, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 9.27, | |
| "grad_norm": 8.859036445617676, | |
| "learning_rate": 1.2994041528833266e-06, | |
| "loss": 0.1902, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 9.45, | |
| "grad_norm": 8.199995994567871, | |
| "learning_rate": 7.323082076153509e-07, | |
| "loss": 0.2104, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 9.64, | |
| "grad_norm": 10.105381965637207, | |
| "learning_rate": 3.2591323257248893e-07, | |
| "loss": 0.2379, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 9.82, | |
| "grad_norm": 14.23947525024414, | |
| "learning_rate": 8.15448036932176e-08, | |
| "loss": 0.217, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 20.24100112915039, | |
| "learning_rate": 0.0, | |
| "loss": 0.2185, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.9399602385685885, | |
| "eval_loss": 0.23698414862155914, | |
| "eval_runtime": 71.0392, | |
| "eval_samples_per_second": 35.403, | |
| "eval_steps_per_second": 1.112, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5500, | |
| "total_flos": 1.301428412334932e+19, | |
| "train_loss": 0.4333542574102228, | |
| "train_runtime": 11179.012, | |
| "train_samples_per_second": 15.727, | |
| "train_steps_per_second": 0.492 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 5500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.301428412334932e+19, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |