roberta-Reflections-goodareas-eval_FeedbackESConv5pp_CARE10pp-sweeps-current

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

• Loss: 0.4328
• Accuracy: 0.8678
• Precision: 0.4273
• Recall: 0.5402
• F1: 0.4772

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: 9.49118803819061e-06
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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	Precision	Recall	F1
0.3453	1.0	80	0.2368	0.8845	0.4444	0.1379	0.2105
0.2686	2.0	160	0.1995	0.8883	0.0	0.0	0.0
0.2467	3.0	240	0.2582	0.8755	0.4561	0.5977	0.5174
0.2346	4.0	320	0.1663	0.9012	0.6923	0.2069	0.3186
0.2153	5.0	400	0.1441	0.9037	0.8	0.1839	0.2991
0.2003	6.0	480	0.2784	0.8267	0.3571	0.6897	0.4706
0.1806	7.0	560	0.1637	0.8999	0.5495	0.5747	0.5618
0.1477	8.0	640	0.2062	0.8639	0.4275	0.6437	0.5138
0.1234	9.0	720	0.2175	0.8626	0.4167	0.5747	0.4831
0.1116	10.0	800	0.1914	0.8845	0.4810	0.4368	0.4578
0.0959	11.0	880	0.3313	0.8485	0.3916	0.6437	0.4870
0.0933	12.0	960	0.3027	0.8575	0.4048	0.5862	0.4789
0.0796	13.0	1040	0.3267	0.8575	0.4032	0.5747	0.4739
0.0688	14.0	1120	0.2958	0.8819	0.4731	0.5057	0.4889
0.0723	15.0	1200	0.4122	0.8575	0.4032	0.5747	0.4739
0.048	16.0	1280	0.5274	0.8447	0.3851	0.6552	0.4851
0.0504	17.0	1360	0.5241	0.8562	0.4031	0.5977	0.4815
0.0353	18.0	1440	0.4845	0.8601	0.4098	0.5747	0.4785
0.0485	19.0	1520	0.5141	0.8562	0.4031	0.5977	0.4815
0.0481	20.0	1600	0.4328	0.8678	0.4273	0.5402	0.4772

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 2.21.0
• Tokenizers 0.21.0