scenario-TCR_data-cardiffnlp_tweet_sentiment_multilingual_all_c

This model is a fine-tuned version of xlm-roberta-base on the tweet_sentiment_multilingual dataset. It achieves the following results on the evaluation set:

• Loss: 2.5108
• Accuracy: 0.6431
• F1: 0.6424

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 134
• 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
0.9471	1.09	500	0.8205	0.6412	0.6387
0.7916	2.17	1000	0.8077	0.6474	0.6462
0.6978	3.26	1500	0.8621	0.6528	0.6534
0.6176	4.35	2000	0.9091	0.6412	0.6363
0.5422	5.43	2500	0.9120	0.6454	0.6440
0.4822	6.52	3000	0.9097	0.6512	0.6469
0.4117	7.61	3500	1.0223	0.6420	0.6406
0.3669	8.7	4000	1.1259	0.6404	0.6427
0.3229	9.78	4500	1.2050	0.6516	0.6489
0.2797	10.87	5000	1.2616	0.6408	0.6415
0.2657	11.96	5500	1.3181	0.6435	0.6412
0.226	13.04	6000	1.4459	0.6400	0.6424
0.2123	14.13	6500	1.5978	0.6389	0.6379
0.1853	15.22	7000	1.6409	0.6412	0.6438
0.1759	16.3	7500	1.6756	0.6485	0.6495
0.1579	17.39	8000	1.6652	0.6412	0.6418
0.1409	18.48	8500	1.9476	0.6389	0.6384
0.1282	19.57	9000	2.0246	0.6285	0.6280
0.1254	20.65	9500	1.9803	0.6412	0.6437
0.1077	21.74	10000	2.0991	0.6447	0.6429
0.097	22.83	10500	2.1971	0.6424	0.6413
0.0965	23.91	11000	2.2161	0.6420	0.6387
0.0859	25.0	11500	2.3387	0.6346	0.6329
0.0744	26.09	12000	2.3921	0.6466	0.6458
0.0693	27.17	12500	2.4696	0.6424	0.6428
0.072	28.26	13000	2.5027	0.6435	0.6431
0.0701	29.35	13500	2.5108	0.6431	0.6424

Framework versions

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