scenario-KD-PO-CDF-ALL-D2_data-cardiffnlp_tweet_sentiment_multilingual_all55

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

• Loss: 3.2137
• Accuracy: 0.6150
• F1: 0.6155

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: 55
• 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
4.3541	1.09	500	3.1863	0.5721	0.5729
3.0722	2.17	1000	3.2433	0.5845	0.5841
2.1971	3.26	1500	3.4852	0.5907	0.5891
1.6717	4.35	2000	3.3784	0.5841	0.5859
1.3144	5.43	2500	3.7616	0.5864	0.5846
1.1448	6.52	3000	3.4276	0.5779	0.5777
0.9875	7.61	3500	3.3781	0.5953	0.5946
0.8762	8.7	4000	3.1471	0.5829	0.5843
0.7963	9.78	4500	3.3721	0.5752	0.5756
0.7369	10.87	5000	3.5433	0.5837	0.5839
0.6786	11.96	5500	3.3185	0.5783	0.5793
0.6271	13.04	6000	3.1903	0.5903	0.5869
0.5889	14.13	6500	3.3031	0.5895	0.5899
0.5569	15.22	7000	3.2597	0.5995	0.5987
0.5229	16.3	7500	3.1515	0.5883	0.5893
0.4985	17.39	8000	3.1954	0.5992	0.5997
0.4751	18.48	8500	3.2670	0.5988	0.5990
0.4643	19.57	9000	3.1884	0.5961	0.5948
0.4434	20.65	9500	3.2196	0.5984	0.5972
0.4381	21.74	10000	3.2211	0.5922	0.5934
0.4193	22.83	10500	3.1172	0.5992	0.5990
0.4113	23.91	11000	3.3000	0.6107	0.6114
0.4008	25.0	11500	3.1179	0.6015	0.6024
0.3914	26.09	12000	3.1496	0.5926	0.5925
0.3861	27.17	12500	3.2896	0.5934	0.5942
0.3768	28.26	13000	3.1737	0.6065	0.6076
0.3767	29.35	13500	3.2205	0.6038	0.6034
0.3687	30.43	14000	3.3223	0.6049	0.6046
0.3663	31.52	14500	3.0866	0.6015	0.6021
0.3542	32.61	15000	3.1709	0.6107	0.6119
0.3549	33.7	15500	3.2819	0.5965	0.5976
0.3476	34.78	16000	3.2686	0.5995	0.6002
0.3452	35.87	16500	3.2558	0.6015	0.6018
0.346	36.96	17000	3.1651	0.5953	0.5962
0.3402	38.04	17500	3.1907	0.6084	0.6086
0.3385	39.13	18000	3.1962	0.6096	0.6098
0.3389	40.22	18500	3.1716	0.6088	0.6091
0.3322	41.3	19000	3.1679	0.6092	0.6096
0.3288	42.39	19500	3.2305	0.6046	0.6044
0.333	43.48	20000	3.2468	0.6011	0.6016
0.33	44.57	20500	3.1011	0.6177	0.6180
0.3322	45.65	21000	3.2199	0.5980	0.5984
0.3267	46.74	21500	3.2237	0.6130	0.6132
0.3268	47.83	22000	3.2034	0.6038	0.6042
0.3238	48.91	22500	3.1129	0.6103	0.6106
0.324	50.0	23000	3.2137	0.6150	0.6155

Framework versions

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