| { | |
| "best_metric": 0.22899456322193146, | |
| "best_model_checkpoint": "output_dir/finetuned_best_qnli/checkpoint-5000", | |
| "epoch": 4.886988393402565, | |
| "global_step": 16000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 1.938912645082468e-05, | |
| "loss": 0.4392, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "eval_accuracy": 0.8553908109097565, | |
| "eval_loss": 0.32711732387542725, | |
| "eval_runtime": 26.9542, | |
| "eval_samples_per_second": 202.677, | |
| "eval_steps_per_second": 12.688, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 1.8778252901649362e-05, | |
| "loss": 0.3536, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "eval_accuracy": 0.8828482518762585, | |
| "eval_loss": 0.2784218490123749, | |
| "eval_runtime": 28.1713, | |
| "eval_samples_per_second": 193.921, | |
| "eval_steps_per_second": 12.14, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 1.816737935247404e-05, | |
| "loss": 0.3288, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_accuracy": 0.8934651290499726, | |
| "eval_loss": 0.2534938454627991, | |
| "eval_runtime": 24.2891, | |
| "eval_samples_per_second": 224.916, | |
| "eval_steps_per_second": 14.08, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.755650580329872e-05, | |
| "loss": 0.3105, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "eval_accuracy": 0.9033498077979132, | |
| "eval_loss": 0.24962939321994781, | |
| "eval_runtime": 24.0235, | |
| "eval_samples_per_second": 227.402, | |
| "eval_steps_per_second": 14.236, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 1.69456322541234e-05, | |
| "loss": 0.3061, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "eval_accuracy": 0.900604063701263, | |
| "eval_loss": 0.24303048849105835, | |
| "eval_runtime": 24.6661, | |
| "eval_samples_per_second": 221.478, | |
| "eval_steps_per_second": 13.865, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 1.6334758704948076e-05, | |
| "loss": 0.2907, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "eval_accuracy": 0.9029837085850265, | |
| "eval_loss": 0.2454785704612732, | |
| "eval_runtime": 25.5666, | |
| "eval_samples_per_second": 213.677, | |
| "eval_steps_per_second": 13.377, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 1.5723885155772757e-05, | |
| "loss": 0.2484, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "eval_accuracy": 0.9125022881200805, | |
| "eval_loss": 0.24116584658622742, | |
| "eval_runtime": 25.7855, | |
| "eval_samples_per_second": 211.864, | |
| "eval_steps_per_second": 13.263, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 1.5113011606597437e-05, | |
| "loss": 0.1937, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "eval_accuracy": 0.9077429983525536, | |
| "eval_loss": 0.25882598757743835, | |
| "eval_runtime": 24.3945, | |
| "eval_samples_per_second": 223.944, | |
| "eval_steps_per_second": 14.02, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 1.4502138057422115e-05, | |
| "loss": 0.1948, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "eval_accuracy": 0.9134175361522973, | |
| "eval_loss": 0.24079638719558716, | |
| "eval_runtime": 25.609, | |
| "eval_samples_per_second": 213.323, | |
| "eval_steps_per_second": 13.355, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "learning_rate": 1.3891264508246794e-05, | |
| "loss": 0.1955, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "eval_accuracy": 0.9126853377265239, | |
| "eval_loss": 0.22899456322193146, | |
| "eval_runtime": 25.5564, | |
| "eval_samples_per_second": 213.763, | |
| "eval_steps_per_second": 13.382, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "learning_rate": 1.3280390959071474e-05, | |
| "loss": 0.1869, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "eval_accuracy": 0.9084751967783269, | |
| "eval_loss": 0.2446000874042511, | |
| "eval_runtime": 24.7359, | |
| "eval_samples_per_second": 220.853, | |
| "eval_steps_per_second": 13.826, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 1.2669517409896153e-05, | |
| "loss": 0.1883, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "eval_accuracy": 0.9115870400878638, | |
| "eval_loss": 0.23336854577064514, | |
| "eval_runtime": 26.4848, | |
| "eval_samples_per_second": 206.269, | |
| "eval_steps_per_second": 12.913, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 1.2058643860720831e-05, | |
| "loss": 0.1897, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "eval_accuracy": 0.9084751967783269, | |
| "eval_loss": 0.2398270219564438, | |
| "eval_runtime": 26.0879, | |
| "eval_samples_per_second": 209.407, | |
| "eval_steps_per_second": 13.11, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 1.1447770311545512e-05, | |
| "loss": 0.1169, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "eval_accuracy": 0.9123192385136372, | |
| "eval_loss": 0.30561476945877075, | |
| "eval_runtime": 24.4434, | |
| "eval_samples_per_second": 223.496, | |
| "eval_steps_per_second": 13.992, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 1.083689676237019e-05, | |
| "loss": 0.1069, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "eval_accuracy": 0.9130514369394106, | |
| "eval_loss": 0.295697420835495, | |
| "eval_runtime": 25.6997, | |
| "eval_samples_per_second": 212.571, | |
| "eval_steps_per_second": 13.308, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 1.0226023213194869e-05, | |
| "loss": 0.1079, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "eval_accuracy": 0.9146988833974007, | |
| "eval_loss": 0.31733793020248413, | |
| "eval_runtime": 25.5822, | |
| "eval_samples_per_second": 213.547, | |
| "eval_steps_per_second": 13.369, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 9.615149664019549e-06, | |
| "loss": 0.1132, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "eval_accuracy": 0.9137836353651839, | |
| "eval_loss": 0.3351670503616333, | |
| "eval_runtime": 26.0424, | |
| "eval_samples_per_second": 209.774, | |
| "eval_steps_per_second": 13.132, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 9.004276114844227e-06, | |
| "loss": 0.1141, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "eval_accuracy": 0.9139666849716274, | |
| "eval_loss": 0.3112892210483551, | |
| "eval_runtime": 25.7504, | |
| "eval_samples_per_second": 212.152, | |
| "eval_steps_per_second": 13.281, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 8.393402565668908e-06, | |
| "loss": 0.118, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "eval_accuracy": 0.9178107267069375, | |
| "eval_loss": 0.28728175163269043, | |
| "eval_runtime": 24.7133, | |
| "eval_samples_per_second": 221.055, | |
| "eval_steps_per_second": 13.839, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 3.05, | |
| "learning_rate": 7.782529016493586e-06, | |
| "loss": 0.0918, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.05, | |
| "eval_accuracy": 0.9145158337909574, | |
| "eval_loss": 0.4170892536640167, | |
| "eval_runtime": 26.16, | |
| "eval_samples_per_second": 208.83, | |
| "eval_steps_per_second": 13.073, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.21, | |
| "learning_rate": 7.171655467318266e-06, | |
| "loss": 0.0678, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 3.21, | |
| "eval_accuracy": 0.9112209408749771, | |
| "eval_loss": 0.4185065031051636, | |
| "eval_runtime": 24.3588, | |
| "eval_samples_per_second": 224.272, | |
| "eval_steps_per_second": 14.04, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "learning_rate": 6.560781918142944e-06, | |
| "loss": 0.0678, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "eval_accuracy": 0.9112209408749771, | |
| "eval_loss": 0.4082688093185425, | |
| "eval_runtime": 24.7254, | |
| "eval_samples_per_second": 220.947, | |
| "eval_steps_per_second": 13.832, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 5.949908368967624e-06, | |
| "loss": 0.0732, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "eval_accuracy": 0.9146988833974007, | |
| "eval_loss": 0.39638856053352356, | |
| "eval_runtime": 25.9223, | |
| "eval_samples_per_second": 210.746, | |
| "eval_steps_per_second": 13.193, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 5.339034819792304e-06, | |
| "loss": 0.0747, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "eval_accuracy": 0.9154310818231741, | |
| "eval_loss": 0.39597952365875244, | |
| "eval_runtime": 25.8782, | |
| "eval_samples_per_second": 211.104, | |
| "eval_steps_per_second": 13.216, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 3.82, | |
| "learning_rate": 4.728161270616982e-06, | |
| "loss": 0.0694, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 3.82, | |
| "eval_accuracy": 0.9167124290682775, | |
| "eval_loss": 0.3776060938835144, | |
| "eval_runtime": 25.8946, | |
| "eval_samples_per_second": 210.97, | |
| "eval_steps_per_second": 13.207, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 4.117287721441662e-06, | |
| "loss": 0.0711, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "eval_accuracy": 0.9154310818231741, | |
| "eval_loss": 0.3837811052799225, | |
| "eval_runtime": 24.4925, | |
| "eval_samples_per_second": 223.048, | |
| "eval_steps_per_second": 13.963, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 3.506414172266341e-06, | |
| "loss": 0.0473, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "eval_accuracy": 0.9146988833974007, | |
| "eval_loss": 0.44068974256515503, | |
| "eval_runtime": 26.0081, | |
| "eval_samples_per_second": 210.05, | |
| "eval_steps_per_second": 13.15, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 4.28, | |
| "learning_rate": 2.8955406230910206e-06, | |
| "loss": 0.043, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 4.28, | |
| "eval_accuracy": 0.9168954786747209, | |
| "eval_loss": 0.4517548680305481, | |
| "eval_runtime": 25.7888, | |
| "eval_samples_per_second": 211.836, | |
| "eval_steps_per_second": 13.262, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 4.43, | |
| "learning_rate": 2.2846670739156996e-06, | |
| "loss": 0.0459, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 4.43, | |
| "eval_accuracy": 0.917993776313381, | |
| "eval_loss": 0.43741610646247864, | |
| "eval_runtime": 28.8351, | |
| "eval_samples_per_second": 189.456, | |
| "eval_steps_per_second": 11.861, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 4.58, | |
| "learning_rate": 1.6737935247403788e-06, | |
| "loss": 0.043, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 4.58, | |
| "eval_accuracy": 0.9154310818231741, | |
| "eval_loss": 0.4457714259624481, | |
| "eval_runtime": 24.7472, | |
| "eval_samples_per_second": 220.752, | |
| "eval_steps_per_second": 13.82, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "learning_rate": 1.0629199755650582e-06, | |
| "loss": 0.0472, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "eval_accuracy": 0.9183598755262676, | |
| "eval_loss": 0.43522849678993225, | |
| "eval_runtime": 25.0601, | |
| "eval_samples_per_second": 217.996, | |
| "eval_steps_per_second": 13.647, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "learning_rate": 4.5204642638973736e-07, | |
| "loss": 0.041, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "eval_accuracy": 0.9183598755262676, | |
| "eval_loss": 0.44270119071006775, | |
| "eval_runtime": 24.5695, | |
| "eval_samples_per_second": 222.349, | |
| "eval_steps_per_second": 13.92, | |
| "step": 16000 | |
| } | |
| ], | |
| "max_steps": 16370, | |
| "num_train_epochs": 5, | |
| "total_flos": 3.3671637309696e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |