scenario-TCR-4_data-en-cardiff_eng_only4

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

• Loss: 3.8804
• Accuracy: 0.5560
• F1: 0.5592

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 44
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.03	60	1.0063	0.5097	0.4877
No log	2.07	120	0.9908	0.5573	0.5382
No log	3.1	180	1.0407	0.5511	0.5519
No log	4.14	240	1.1788	0.5714	0.5690
No log	5.17	300	1.3054	0.5670	0.5687
No log	6.21	360	1.5511	0.5569	0.5582
No log	7.24	420	1.7923	0.5428	0.5423
No log	8.28	480	1.8397	0.5534	0.5560
0.5939	9.31	540	2.2461	0.5357	0.5376
0.5939	10.34	600	2.7671	0.5441	0.5465
0.5939	11.38	660	2.8160	0.5392	0.5395
0.5939	12.41	720	3.0003	0.5348	0.5351
0.5939	13.45	780	3.2829	0.5357	0.5382
0.5939	14.48	840	3.2854	0.5357	0.5371
0.5939	15.52	900	3.0484	0.5428	0.5464
0.5939	16.55	960	3.2120	0.5511	0.5520
0.0837	17.59	1020	3.3202	0.5481	0.5503
0.0837	18.62	1080	3.5737	0.5251	0.5212
0.0837	19.66	1140	3.5754	0.5432	0.5440
0.0837	20.69	1200	3.8077	0.5304	0.5332
0.0837	21.72	1260	3.7584	0.5414	0.5440
0.0837	22.76	1320	3.7424	0.5498	0.5519
0.0837	23.79	1380	3.7333	0.5595	0.5619
0.0837	24.83	1440	3.8425	0.5542	0.5579
0.0088	25.86	1500	3.8197	0.5591	0.5618
0.0088	26.9	1560	4.0462	0.5392	0.5411
0.0088	27.93	1620	3.9146	0.5520	0.5548
0.0088	28.97	1680	3.8703	0.5573	0.5605
0.0088	30.0	1740	3.8804	0.5560	0.5592

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3