stance_class_l

This model is a fine-tuned version of vinai/bertweet-base on the dataset of 804 labeled tweets on the cancer risk controversy of Roundup Weedkiller . It classified the stance of an individual's tweet toward Bayer, Monsanto, or other relevant organizations in the crisis. Two stances are classified: (0) Aggressive, (1) Non-Aggressive (neutral and accommodative). It achieves the following results on the evaluation set:

• Loss: 0.6084
• Accuracy: 0.8447

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: 2.924e-05
• train_batch_size: 30
• eval_batch_size: 30
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.3566	1.0	17	0.4855	0.7578
0.2532	2.0	34	0.3632	0.8509
0.2351	3.0	51	0.3773	0.8509
0.043	4.0	68	0.3553	0.8571
0.08	5.0	85	0.4682	0.8447
0.3089	6.0	102	0.4686	0.8509
0.035	7.0	119	0.5876	0.8323
0.0188	8.0	136	0.5469	0.8571
0.021	9.0	153	0.5022	0.8447
0.0533	10.0	170	0.5240	0.8385
0.0175	11.0	187	0.6352	0.8447
0.0106	12.0	204	0.5856	0.8447
1.9534	13.0	221	0.5938	0.8509
0.0143	14.0	238	0.6074	0.8447
0.0079	15.0	255	0.6084	0.8447

Framework versions

• Transformers 4.24.0
• Pytorch 1.13.0+cu117
• Datasets 2.12.0
• Tokenizers 0.13.2