scenario-KD-PO-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_all_gamma

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

• Loss: 3.5245
• Accuracy: 0.5517
• F1: 0.5524

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: 88458
• 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
4.8673	1.09	500	4.1044	0.4356	0.4381
4.0355	2.17	1000	3.8162	0.5004	0.4829
3.4812	3.26	1500	3.3484	0.5312	0.5299
3.1323	4.35	2000	3.3401	0.5502	0.5500
2.7632	5.43	2500	3.5126	0.5471	0.5441
2.5101	6.52	3000	3.5161	0.5444	0.5412
2.3266	7.61	3500	3.6769	0.5367	0.5263
2.1096	8.7	4000	3.6299	0.5513	0.5501
1.972	9.78	4500	3.4289	0.5432	0.5428
1.8345	10.87	5000	3.3890	0.5502	0.5464
1.711	11.96	5500	3.3365	0.5548	0.5553
1.6043	13.04	6000	3.4657	0.5529	0.5527
1.4994	14.13	6500	3.3948	0.5494	0.5500
1.404	15.22	7000	3.5906	0.5529	0.5533
1.3423	16.3	7500	3.5538	0.5575	0.5555
1.2991	17.39	8000	3.5762	0.5532	0.5539
1.217	18.48	8500	3.6649	0.5517	0.5518
1.1763	19.57	9000	3.5238	0.5513	0.5503
1.1249	20.65	9500	3.5218	0.5436	0.5453
1.0774	21.74	10000	3.7103	0.5617	0.5622
1.0558	22.83	10500	3.6698	0.5567	0.5558
1.0036	23.91	11000	3.4754	0.5648	0.5645
0.9734	25.0	11500	3.5782	0.5490	0.5483
0.9614	26.09	12000	3.4920	0.5586	0.5600
0.9221	27.17	12500	3.5416	0.5440	0.5436
0.905	28.26	13000	3.5065	0.5640	0.5635
0.8845	29.35	13500	3.6653	0.5463	0.5464
0.8614	30.43	14000	3.5104	0.5583	0.5571
0.8414	31.52	14500	3.6002	0.5548	0.5554
0.8328	32.61	15000	3.5431	0.5544	0.5527
0.8134	33.7	15500	3.5080	0.5590	0.5585
0.7973	34.78	16000	3.4150	0.5583	0.5578
0.7887	35.87	16500	3.6270	0.5486	0.5502
0.7778	36.96	17000	3.6464	0.5494	0.5491
0.7662	38.04	17500	3.5100	0.5633	0.5627
0.7553	39.13	18000	3.5580	0.5532	0.5537
0.7426	40.22	18500	3.4555	0.5594	0.5583
0.7494	41.3	19000	3.5871	0.5590	0.5554
0.7252	42.39	19500	3.4094	0.5590	0.5595
0.7293	43.48	20000	3.4817	0.5656	0.5661
0.7103	44.57	20500	3.4964	0.5594	0.5596
0.718	45.65	21000	3.4770	0.5598	0.5593
0.7147	46.74	21500	3.4938	0.5613	0.5616
0.7014	47.83	22000	3.4664	0.5571	0.5567
0.6991	48.91	22500	3.4357	0.5606	0.5606
0.6944	50.0	23000	3.5245	0.5517	0.5524

Framework versions

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