t5_cause_classifier

This model is a fine-tuned version of google/flan-t5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2104
• F1: 0.8112
• Accuracy: 0.3196

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: 32
• eval_batch_size: 8
• 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	124	0.3976	0.5740	0.0
No log	2.0	248	0.2948	0.7116	0.0907
No log	3.0	372	0.2657	0.7396	0.1200
No log	4.0	496	0.2514	0.7599	0.1724
0.3526	5.0	620	0.2406	0.7706	0.2006
0.3526	6.0	744	0.2318	0.7809	0.2147
0.3526	7.0	868	0.2267	0.7910	0.2591
0.3526	8.0	992	0.2227	0.7949	0.2742
0.2263	9.0	1116	0.2178	0.7993	0.2772
0.2263	10.0	1240	0.2161	0.8028	0.2923
0.2263	11.0	1364	0.2158	0.8024	0.2873
0.2263	12.0	1488	0.2144	0.8033	0.2984
0.2005	13.0	1612	0.2125	0.8074	0.3054
0.2005	14.0	1736	0.2118	0.8076	0.3206
0.2005	15.0	1860	0.2120	0.8093	0.3095
0.2005	16.0	1984	0.2122	0.8088	0.3196
0.1877	17.0	2108	0.2109	0.8096	0.3155
0.1877	18.0	2232	0.2106	0.8103	0.3135
0.1877	19.0	2356	0.2109	0.8099	0.3155
0.1877	20.0	2480	0.2104	0.8112	0.3196

Framework versions

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