roberta-base-riddle-finetuned_new

This model is a fine-tuned version of roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3741
• Accuracy: 0.8250

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: 0.0005
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	12	1.3036	0.4500
No log	2.0	24	1.1112	0.4750
No log	3.0	36	1.0265	0.4500
No log	4.0	48	1.1130	0.6000
No log	5.0	60	0.8603	0.625
No log	6.0	72	0.8099	0.7750
No log	7.0	84	0.8668	0.7250
No log	8.0	96	0.7217	0.8000
No log	9.0	108	0.7143	0.8250
No log	10.0	120	0.6371	0.8250
No log	11.0	132	0.6327	0.7250
No log	12.0	144	0.5974	0.75
No log	13.0	156	0.5160	0.8000
No log	14.0	168	0.5336	0.75
No log	15.0	180	0.5201	0.8000
No log	16.0	192	0.4121	0.8250
No log	17.0	204	0.4145	0.8000
No log	18.0	216	0.4475	0.875
No log	19.0	228	0.4147	0.8250
No log	20.0	240	0.3818	0.7750
No log	21.0	252	0.4136	0.75
No log	22.0	264	0.4364	0.75
No log	23.0	276	0.4180	0.7250
No log	24.0	288	0.4145	0.75
No log	25.0	300	0.4141	0.8000
No log	26.0	312	0.3948	0.8000
No log	27.0	324	0.3930	0.8250
No log	28.0	336	0.3851	0.8250
No log	29.0	348	0.3745	0.8250
No log	30.0	360	0.3741	0.8250

Framework versions

• Transformers 4.37.2
• Pytorch 1.13.1+cu117
• Datasets 2.16.1
• Tokenizers 0.15.0