scenario-TCR_data-cardiffnlp_tweet_sentiment_multilingual_all_b

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: 2.7148
• Accuracy: 0.6373
• F1: 0.6373

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: 6666
• 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
0.9727	1.09	500	0.8537	0.6134	0.6083
0.8035	2.17	1000	0.8204	0.6312	0.6114
0.7087	3.26	1500	0.8422	0.6393	0.6272
0.6246	4.35	2000	0.8451	0.6431	0.6369
0.5428	5.43	2500	0.9453	0.6412	0.6399
0.4829	6.52	3000	1.0229	0.6431	0.6404
0.4161	7.61	3500	0.9831	0.6478	0.6473
0.3612	8.7	4000	1.1288	0.6400	0.6395
0.3086	9.78	4500	1.2783	0.6335	0.6326
0.2779	10.87	5000	1.3253	0.625	0.6229
0.244	11.96	5500	1.4854	0.6404	0.6388
0.2192	13.04	6000	1.5661	0.6308	0.6331
0.1881	14.13	6500	1.7309	0.6304	0.6282
0.1654	15.22	7000	1.8752	0.6316	0.6298
0.162	16.3	7500	1.7038	0.6377	0.6390
0.1392	17.39	8000	1.8729	0.6219	0.6226
0.1234	18.48	8500	1.9495	0.6312	0.6317
0.1125	19.57	9000	2.1487	0.6308	0.6334
0.1035	20.65	9500	2.1859	0.6404	0.6406
0.0988	21.74	10000	2.2385	0.6281	0.6277
0.0837	22.83	10500	2.3919	0.6377	0.6383
0.0793	23.91	11000	2.5023	0.6408	0.6420
0.0773	25.0	11500	2.5799	0.6319	0.6316
0.0724	26.09	12000	2.6326	0.6327	0.6335
0.0659	27.17	12500	2.7253	0.6335	0.6336
0.0589	28.26	13000	2.7358	0.6343	0.6345
0.0613	29.35	13500	2.7148	0.6373	0.6373

Framework versions

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