scenario-NON-KD-PO-COPY-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_

This model is a fine-tuned version of haryoaw/scenario-MDBT-TCR-TSM on the tweet_sentiment_multilingual dataset. It achieves the following results on the evaluation set:

• Loss: 4.8189
• Accuracy: 0.5498
• F1: 0.5496

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: 66
• 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
1.063	1.0870	500	1.0629	0.4938	0.4516
0.9517	2.1739	1000	1.0992	0.5239	0.5103
0.8617	3.2609	1500	1.0416	0.5413	0.5397
0.75	4.3478	2000	1.1450	0.5463	0.5395
0.6494	5.4348	2500	1.2684	0.5548	0.5469
0.5512	6.5217	3000	1.5424	0.5390	0.5255
0.4603	7.6087	3500	1.5437	0.5471	0.5463
0.3867	8.6957	4000	1.4993	0.5536	0.5513
0.3247	9.7826	4500	1.7700	0.5444	0.5410
0.2599	10.8696	5000	1.8467	0.5509	0.5510
0.2228	11.9565	5500	2.0318	0.5421	0.5426
0.187	13.0435	6000	2.3511	0.5448	0.5448
0.1685	14.1304	6500	2.6019	0.5502	0.5499
0.1471	15.2174	7000	2.7888	0.5421	0.5410
0.1337	16.3043	7500	2.7355	0.5448	0.5416
0.1279	17.3913	8000	2.6407	0.5475	0.5385
0.1073	18.4783	8500	3.0122	0.5532	0.5511
0.1081	19.5652	9000	3.1280	0.5532	0.5524
0.0998	20.6522	9500	2.8679	0.5448	0.5461
0.0835	21.7391	10000	3.4452	0.5459	0.5461
0.085	22.8261	10500	3.2664	0.5370	0.5366
0.0744	23.9130	11000	3.8876	0.5374	0.5378
0.0704	25.0	11500	3.6019	0.5417	0.5388
0.0648	26.0870	12000	3.7721	0.5463	0.5471
0.0589	27.1739	12500	4.0283	0.5421	0.5418
0.0593	28.2609	13000	4.1378	0.5390	0.5359
0.052	29.3478	13500	4.0042	0.5378	0.5309
0.0452	30.4348	14000	4.5220	0.5305	0.5316
0.0452	31.5217	14500	4.1742	0.5336	0.5328
0.0391	32.6087	15000	4.2415	0.5432	0.5428
0.0376	33.6957	15500	3.7389	0.5505	0.5495
0.0329	34.7826	16000	4.3008	0.5529	0.5527
0.0323	35.8696	16500	4.2965	0.5467	0.5431
0.0293	36.9565	17000	4.2941	0.5424	0.5403
0.0283	38.0435	17500	4.4543	0.5502	0.5512
0.0258	39.1304	18000	4.4241	0.5405	0.5391
0.0201	40.2174	18500	4.7437	0.5351	0.5354
0.0282	41.3043	19000	4.2344	0.5475	0.5462
0.0223	42.3913	19500	4.4261	0.5463	0.5475
0.0188	43.4783	20000	4.6276	0.5463	0.5429
0.0128	44.5652	20500	4.9760	0.5505	0.5497
0.0166	45.6522	21000	4.7753	0.5459	0.5442
0.0166	46.7391	21500	4.7611	0.5513	0.5511
0.0137	47.8261	22000	4.7747	0.5532	0.5530
0.0143	48.9130	22500	4.8316	0.5494	0.5492
0.0143	50.0	23000	4.8189	0.5498	0.5496

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1