absa-train-service-roberta-large

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: 0.8683
• Accuracy: 0.7424
• Precision: 0.7345
• Recall: 0.7367
• F1: 0.7302

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.2255	1.0	469	2.0677	0.3296	0.1937	0.3250	0.2297
1.8236	2.0	938	1.7061	0.504	0.5413	0.4914	0.4567
1.5384	3.0	1407	1.4381	0.552	0.5944	0.5549	0.5196
1.4301	4.0	1876	1.3316	0.5984	0.6000	0.5990	0.5618
1.3776	5.0	2345	1.1645	0.6576	0.6817	0.6491	0.6332
1.2078	6.0	2814	1.0967	0.6448	0.7035	0.6348	0.6110
1.2535	7.0	3283	1.0565	0.7008	0.7467	0.6967	0.7066
1.2921	8.0	3752	1.0049	0.6976	0.7013	0.6884	0.6813
1.178	9.0	4221	1.0438	0.648	0.7746	0.6423	0.6387
1.2324	10.0	4690	1.0203	0.6896	0.7096	0.6831	0.6704
1.1899	11.0	5159	1.0193	0.6864	0.7391	0.6819	0.6834
1.1515	12.0	5628	0.9722	0.6944	0.7164	0.6924	0.6860
1.1604	13.0	6097	0.9372	0.7312	0.7543	0.7311	0.7259
1.1229	14.0	6566	0.9265	0.72	0.7278	0.7139	0.7147
1.1459	15.0	7035	0.8896	0.7376	0.7264	0.7323	0.7183
1.1281	16.0	7504	0.9074	0.7152	0.7107	0.7087	0.7012
1.1794	17.0	7973	0.8914	0.7424	0.7293	0.7354	0.7266
1.1101	18.0	8442	0.8707	0.7216	0.7161	0.7141	0.7059
1.1215	19.0	8911	0.8656	0.7408	0.7322	0.7348	0.7274
1.0483	20.0	9380	0.8683	0.7424	0.7345	0.7367	0.7302

Framework versions

• Transformers 4.43.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1