| { | |
| "best_metric": 0.7659574468085106, | |
| "best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-2/checkpoint-112", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 160, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.8302626013755798, | |
| "learning_rate": 9.517539799539297e-06, | |
| "loss": 0.6966, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 0.6829280257225037, | |
| "learning_rate": 1.9035079599078593e-05, | |
| "loss": 0.6813, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 0.5496665835380554, | |
| "learning_rate": 2.8552619398617893e-05, | |
| "loss": 0.6451, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_f1": 0.7326732673267327, | |
| "eval_loss": 0.6860618591308594, | |
| "eval_runtime": 1.3401, | |
| "eval_samples_per_second": 47.759, | |
| "eval_steps_per_second": 5.97, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 1.3569624423980713, | |
| "learning_rate": 2.9610123820788925e-05, | |
| "loss": 0.6113, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 0.3838456869125366, | |
| "learning_rate": 2.8552619398617893e-05, | |
| "loss": 0.5997, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 0.46842828392982483, | |
| "learning_rate": 2.7495114976446858e-05, | |
| "loss": 0.6133, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_f1": 0.7326732673267327, | |
| "eval_loss": 0.719573974609375, | |
| "eval_runtime": 1.3426, | |
| "eval_samples_per_second": 47.668, | |
| "eval_steps_per_second": 5.958, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 0.7610273957252502, | |
| "learning_rate": 2.6437610554275826e-05, | |
| "loss": 0.5971, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 0.5527757406234741, | |
| "learning_rate": 2.5380106132104794e-05, | |
| "loss": 0.6373, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 0.9699375033378601, | |
| "learning_rate": 2.4322601709933762e-05, | |
| "loss": 0.5701, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_f1": 0.7326732673267327, | |
| "eval_loss": 0.6987495422363281, | |
| "eval_runtime": 1.344, | |
| "eval_samples_per_second": 47.619, | |
| "eval_steps_per_second": 5.952, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "grad_norm": 0.5822675824165344, | |
| "learning_rate": 2.3265097287762727e-05, | |
| "loss": 0.5492, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 3.44, | |
| "grad_norm": 0.9859365224838257, | |
| "learning_rate": 2.2207592865591692e-05, | |
| "loss": 0.5552, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 0.7946156859397888, | |
| "learning_rate": 2.115008844342066e-05, | |
| "loss": 0.5459, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_f1": 0.7326732673267327, | |
| "eval_loss": 0.6806888580322266, | |
| "eval_runtime": 1.3463, | |
| "eval_samples_per_second": 47.538, | |
| "eval_steps_per_second": 5.942, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 4.06, | |
| "grad_norm": 1.2706314325332642, | |
| "learning_rate": 2.009258402124963e-05, | |
| "loss": 0.5792, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 4.38, | |
| "grad_norm": 1.456335186958313, | |
| "learning_rate": 1.9035079599078593e-05, | |
| "loss": 0.4967, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 4.69, | |
| "grad_norm": 1.6856915950775146, | |
| "learning_rate": 1.797757517690756e-05, | |
| "loss": 0.4966, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 4.080612659454346, | |
| "learning_rate": 1.692007075473653e-05, | |
| "loss": 0.4589, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_f1": 0.7474747474747475, | |
| "eval_loss": 0.6495094299316406, | |
| "eval_runtime": 1.348, | |
| "eval_samples_per_second": 47.477, | |
| "eval_steps_per_second": 5.935, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 5.31, | |
| "grad_norm": 2.8024661540985107, | |
| "learning_rate": 1.5862566332565498e-05, | |
| "loss": 0.4312, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 5.62, | |
| "grad_norm": 3.024564027786255, | |
| "learning_rate": 1.4805061910394462e-05, | |
| "loss": 0.4318, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 5.94, | |
| "grad_norm": 3.236055374145508, | |
| "learning_rate": 1.3747557488223429e-05, | |
| "loss": 0.3968, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_f1": 0.7391304347826088, | |
| "eval_loss": 0.6395978927612305, | |
| "eval_runtime": 1.3605, | |
| "eval_samples_per_second": 47.043, | |
| "eval_steps_per_second": 5.88, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "grad_norm": 2.6648523807525635, | |
| "learning_rate": 1.2690053066052397e-05, | |
| "loss": 0.3623, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 6.56, | |
| "grad_norm": 2.6884357929229736, | |
| "learning_rate": 1.1632548643881364e-05, | |
| "loss": 0.4061, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 6.88, | |
| "grad_norm": 3.5860610008239746, | |
| "learning_rate": 1.057504422171033e-05, | |
| "loss": 0.3356, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_f1": 0.7659574468085106, | |
| "eval_loss": 0.7118406295776367, | |
| "eval_runtime": 1.3559, | |
| "eval_samples_per_second": 47.202, | |
| "eval_steps_per_second": 5.9, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 7.19, | |
| "grad_norm": 3.730743169784546, | |
| "learning_rate": 9.517539799539297e-06, | |
| "loss": 0.298, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "grad_norm": 5.971029281616211, | |
| "learning_rate": 8.460035377368265e-06, | |
| "loss": 0.3183, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 7.81, | |
| "grad_norm": 3.3043053150177, | |
| "learning_rate": 7.402530955197231e-06, | |
| "loss": 0.2788, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_f1": 0.7272727272727273, | |
| "eval_loss": 0.6860837936401367, | |
| "eval_runtime": 1.3424, | |
| "eval_samples_per_second": 47.676, | |
| "eval_steps_per_second": 5.96, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 8.12, | |
| "grad_norm": 3.310666084289551, | |
| "learning_rate": 6.556527417460405e-06, | |
| "loss": 0.3052, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 8.44, | |
| "grad_norm": 3.9955413341522217, | |
| "learning_rate": 5.499022995289372e-06, | |
| "loss": 0.2833, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 8.75, | |
| "grad_norm": 2.4762678146362305, | |
| "learning_rate": 4.441518573118339e-06, | |
| "loss": 0.2387, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_f1": 0.7333333333333333, | |
| "eval_loss": 0.7196941375732422, | |
| "eval_runtime": 1.3575, | |
| "eval_samples_per_second": 47.147, | |
| "eval_steps_per_second": 5.893, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 9.06, | |
| "grad_norm": 5.2802581787109375, | |
| "learning_rate": 3.3840141509473056e-06, | |
| "loss": 0.3212, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 9.38, | |
| "grad_norm": 4.355774402618408, | |
| "learning_rate": 2.326509728776273e-06, | |
| "loss": 0.2311, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 9.69, | |
| "grad_norm": 3.7583141326904297, | |
| "learning_rate": 1.2690053066052396e-06, | |
| "loss": 0.2676, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 3.6632018089294434, | |
| "learning_rate": 2.115008844342066e-07, | |
| "loss": 0.254, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_f1": 0.7472527472527473, | |
| "eval_loss": 0.7351865768432617, | |
| "eval_runtime": 1.353, | |
| "eval_samples_per_second": 47.304, | |
| "eval_steps_per_second": 5.913, | |
| "step": 160 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 160, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 8990959054911696.0, | |
| "train_batch_size": 12, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 3.045612735852575e-05, | |
| "per_device_train_batch_size": 12 | |
| } | |
| } | |