twitter-roberta-base-sentiment-latest-finetuned-FG-SINGLE_SENTENCE-NEWS

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

• Loss: 3.2822
• Accuracy: 0.6305
• F1: 0.6250

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: 6e-05
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	321	0.9646	0.5624	0.4048
0.9537	2.0	642	0.9474	0.5644	0.4176
0.9537	3.0	963	0.9008	0.5903	0.5240
0.858	4.0	1284	0.9939	0.5999	0.5846
0.5908	5.0	1605	1.0947	0.6108	0.6026
0.5908	6.0	1926	1.2507	0.5740	0.5823
0.3676	7.0	2247	1.4717	0.6128	0.6017
0.2246	8.0	2568	1.6726	0.5965	0.6003
0.2246	9.0	2889	1.8041	0.6380	0.6298
0.1468	10.0	3210	1.9796	0.6053	0.6026
0.1161	11.0	3531	2.0988	0.6237	0.6202
0.1161	12.0	3852	2.4171	0.5944	0.5989
0.0916	13.0	4173	2.3326	0.6374	0.6288
0.0916	14.0	4494	2.5472	0.6360	0.6245
0.0661	15.0	4815	2.9127	0.6176	0.6187
0.0454	16.0	5136	2.9133	0.6326	0.6276
0.0454	17.0	5457	3.1299	0.6210	0.6162
0.0337	18.0	5778	3.1828	0.6224	0.6188
0.0223	19.0	6099	3.2655	0.6299	0.6223
0.0223	20.0	6420	3.2822	0.6305	0.6250

Framework versions

• Transformers 4.16.2
• Pytorch 1.9.1
• Datasets 1.18.4
• Tokenizers 0.11.6