roberta-base_fold_1_binary_v1

This model is a fine-tuned version of roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.4984
• F1: 0.8339

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	288	0.3819	0.8117
0.4108	2.0	576	0.3696	0.8281
0.4108	3.0	864	0.4890	0.8343
0.2261	4.0	1152	0.7605	0.8298
0.2261	5.0	1440	0.7754	0.8307
0.1404	6.0	1728	0.7650	0.8174
0.0962	7.0	2016	0.8539	0.8315
0.0962	8.0	2304	1.0770	0.8263
0.0433	9.0	2592	1.1450	0.8292
0.0433	10.0	2880	1.1700	0.8205
0.0344	11.0	3168	1.2376	0.8241
0.0344	12.0	3456	1.2688	0.8329
0.0219	13.0	3744	1.3276	0.8283
0.0123	14.0	4032	1.2930	0.8320
0.0123	15.0	4320	1.4631	0.8266
0.0177	16.0	4608	1.4326	0.8270
0.0177	17.0	4896	1.4770	0.8334
0.0053	18.0	5184	1.5972	0.8214
0.0053	19.0	5472	1.5331	0.8327
0.0045	20.0	5760	1.5487	0.8359
0.0086	21.0	6048	1.4610	0.8315
0.0086	22.0	6336	1.4685	0.8353
0.0071	23.0	6624	1.4933	0.8358
0.0071	24.0	6912	1.4898	0.8310
0.0022	25.0	7200	1.4984	0.8339

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1