Predicts if your motto makes you an Autobot or Decepticon!

roberta-base-finetuned-TF-mottos

This model is a fine-tuned version of roberta-base on all of the Autobot and Decepticon mottos from 1984 - 1986. It achieves the following results on the evaluation set:

• Loss: 1.4868
• Accuracy: 0.7586
• F1: 0.7580

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• L2: 0.25
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.686	1.0	16	0.6934	0.5172	0.3527
0.681	2.0	32	0.6899	0.5172	0.3527
0.6585	3.0	48	0.6632	0.6552	0.6225
0.5878	4.0	64	0.6862	0.5862	0.5862
0.3434	5.0	80	0.8420	0.6207	0.6152
0.1648	6.0	96	1.3011	0.6207	0.6066
0.0624	7.0	112	1.3493	0.7241	0.7241
0.0487	8.0	128	1.5802	0.6897	0.6851
0.0182	9.0	144	1.4868	0.7586	0.7580
0.0051	10.0	160	2.2575	0.6552	0.6467
0.0261	11.0	176	2.5361	0.6552	0.6467
0.0094	12.0	192	2.1784	0.6897	0.6889
0.0025	13.0	208	2.2300	0.6897	0.6889
0.0006	14.0	224	2.1252	0.6897	0.6889
0.0006	15.0	240	2.2771	0.6897	0.6889
0.0113	16.0	256	2.2596	0.6897	0.6889
0.0007	17.0	272	2.1959	0.6897	0.6889
0.0005	18.0	288	2.2496	0.6897	0.6889
0.0004	19.0	304	2.3294	0.6897	0.6889
0.0039	20.0	320	2.3570	0.6897	0.6889

Framework versions

• Transformers 4.48.2
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0