bart_cause_classifier

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

• Loss: 0.2310
• F1: 0.8139
• Accuracy: 0.4022

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: 64
• eval_batch_size: 64
• 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	F1	Accuracy
No log	1.0	62	0.2882	0.7066	0.0444
No log	2.0	124	0.2538	0.7585	0.1643
No log	3.0	186	0.2338	0.7840	0.2581
No log	4.0	248	0.2202	0.8000	0.2954
No log	5.0	310	0.2218	0.7997	0.3306
No log	6.0	372	0.2146	0.8093	0.3488
No log	7.0	434	0.2157	0.8073	0.3498
No log	8.0	496	0.2146	0.8089	0.3629
0.227	9.0	558	0.2223	0.8086	0.3972
0.227	10.0	620	0.2215	0.8088	0.3639
0.227	11.0	682	0.2191	0.8160	0.3982
0.227	12.0	744	0.2227	0.8119	0.3821
0.227	13.0	806	0.2293	0.8070	0.3790
0.227	14.0	868	0.2310	0.8089	0.3780
0.227	15.0	930	0.2267	0.8125	0.3982
0.227	16.0	992	0.2275	0.8116	0.3851
0.1282	17.0	1054	0.2289	0.8150	0.4052
0.1282	18.0	1116	0.2320	0.8106	0.3891
0.1282	19.0	1178	0.2321	0.8114	0.3942
0.1282	20.0	1240	0.2310	0.8139	0.4022

Framework versions

• Transformers 4.41.1
• Pytorch 1.13.1+cu117
• Datasets 2.19.1
• Tokenizers 0.19.1