scenario-KD-PR-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_all_delta

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

• Loss: 3.4263
• Accuracy: 0.5617
• F1: 0.5621

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: 7777
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
4.8568	1.09	500	3.9914	0.4734	0.4688
3.9413	2.17	1000	3.8048	0.5127	0.5070
3.4502	3.26	1500	3.5184	0.5289	0.5171
3.0935	4.35	2000	3.3541	0.5436	0.5418
2.7635	5.43	2500	3.3827	0.5444	0.5443
2.5494	6.52	3000	3.4817	0.5428	0.5440
2.3131	7.61	3500	3.3051	0.5640	0.5567
2.131	8.7	4000	3.2511	0.5548	0.5574
1.9564	9.78	4500	3.4609	0.5583	0.5544
1.8216	10.87	5000	3.2391	0.5502	0.5520
1.7048	11.96	5500	3.2188	0.5525	0.5531
1.575	13.04	6000	3.2912	0.5637	0.5611
1.4693	14.13	6500	3.5853	0.5629	0.5628
1.4236	15.22	7000	3.3838	0.5421	0.5441
1.3372	16.3	7500	3.5262	0.5590	0.5583
1.2829	17.39	8000	3.6001	0.5552	0.5535
1.2351	18.48	8500	3.3745	0.5525	0.5513
1.1562	19.57	9000	3.3239	0.5706	0.5726
1.1264	20.65	9500	3.4648	0.5490	0.5507
1.0806	21.74	10000	3.4269	0.5652	0.5652
1.066	22.83	10500	3.3415	0.5613	0.5617
1.0144	23.91	11000	3.5331	0.5610	0.5623
0.9746	25.0	11500	3.5136	0.5625	0.5625
0.9415	26.09	12000	3.5623	0.5540	0.5530
0.932	27.17	12500	3.5626	0.5633	0.5622
0.9016	28.26	13000	3.6071	0.5467	0.5460
0.8847	29.35	13500	3.5201	0.5513	0.5519
0.8713	30.43	14000	3.5412	0.5660	0.5655
0.8459	31.52	14500	3.5206	0.5556	0.5554
0.8255	32.61	15000	3.4715	0.5552	0.5563
0.8082	33.7	15500	3.4875	0.5579	0.5584
0.7899	34.78	16000	3.4935	0.5775	0.5758
0.7958	35.87	16500	3.4224	0.5544	0.5555
0.7745	36.96	17000	3.3893	0.5671	0.5686
0.7666	38.04	17500	3.3972	0.5629	0.5640
0.7574	39.13	18000	3.5453	0.5706	0.5698
0.7468	40.22	18500	3.4342	0.5671	0.5660
0.7449	41.3	19000	3.3906	0.5640	0.5642
0.7338	42.39	19500	3.4109	0.5721	0.5728
0.7157	43.48	20000	3.3499	0.5721	0.5717
0.7285	44.57	20500	3.2780	0.5718	0.5718
0.7101	45.65	21000	3.3873	0.5648	0.5653
0.7144	46.74	21500	3.4731	0.5613	0.5621
0.7158	47.83	22000	3.4394	0.5733	0.5728
0.7016	48.91	22500	3.4609	0.5544	0.5545
0.7055	50.0	23000	3.4263	0.5617	0.5621

Framework versions

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