scenario-KD-PO-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_all_delta

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: 3.3760
• Accuracy: 0.5586
• F1: 0.5590

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
4.88	1.09	500	4.1661	0.4518	0.4462
3.9513	2.17	1000	3.7172	0.5154	0.5102
3.4556	3.26	1500	3.5612	0.5293	0.5221
3.0959	4.35	2000	3.5271	0.5440	0.5349
2.7593	5.43	2500	3.3105	0.5602	0.5595
2.548	6.52	3000	3.4842	0.5355	0.5360
2.3323	7.61	3500	3.2883	0.5575	0.5444
2.1215	8.7	4000	3.2282	0.5583	0.5607
1.9719	9.78	4500	3.5734	0.5594	0.5534
1.8238	10.87	5000	3.2866	0.5552	0.5538
1.703	11.96	5500	3.2345	0.5556	0.5557
1.5931	13.04	6000	3.2219	0.5583	0.5539
1.4959	14.13	6500	3.3073	0.5637	0.5644
1.4198	15.22	7000	3.5221	0.5579	0.5541
1.35	16.3	7500	3.5125	0.5660	0.5643
1.2937	17.39	8000	3.5089	0.5640	0.5637
1.2282	18.48	8500	3.4262	0.5664	0.5658
1.1698	19.57	9000	3.3739	0.5598	0.5593
1.1402	20.65	9500	3.4930	0.5521	0.5541
1.0874	21.74	10000	3.4935	0.5625	0.5602
1.0652	22.83	10500	3.3963	0.5482	0.5478
1.0191	23.91	11000	3.4823	0.5571	0.5583
0.9868	25.0	11500	3.6035	0.5579	0.5586
0.9487	26.09	12000	3.6034	0.5525	0.5488
0.936	27.17	12500	3.5428	0.5556	0.5542
0.9116	28.26	13000	3.6023	0.5532	0.5509
0.8868	29.35	13500	3.5292	0.5579	0.5581
0.8733	30.43	14000	3.4206	0.5594	0.5589
0.8538	31.52	14500	3.4417	0.5594	0.5592
0.8289	32.61	15000	3.4970	0.5579	0.5584
0.823	33.7	15500	3.4860	0.5625	0.5617
0.7992	34.78	16000	3.5193	0.5671	0.5659
0.7974	35.87	16500	3.3709	0.5490	0.5497
0.7775	36.96	17000	3.3854	0.5706	0.5720
0.7691	38.04	17500	3.3827	0.5698	0.5700
0.758	39.13	18000	3.4608	0.5818	0.5818
0.757	40.22	18500	3.3860	0.5683	0.5681
0.7481	41.3	19000	3.3757	0.5687	0.5686
0.7387	42.39	19500	3.4830	0.5714	0.5707
0.7276	43.48	20000	3.3942	0.5617	0.5611
0.7279	44.57	20500	3.3357	0.5725	0.5726
0.7127	45.65	21000	3.3856	0.5521	0.5523
0.7148	46.74	21500	3.4401	0.5660	0.5673
0.7219	47.83	22000	3.4684	0.5629	0.5627
0.7005	48.91	22500	3.3840	0.5625	0.5631
0.7114	50.0	23000	3.3760	0.5586	0.5590

Framework versions

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