finetuned_1L_BERT_newcode_5epoch-f1score

This model is a fine-tuned version of Youssef320/LSTM-finetuned-50label-15epoch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.0134
• Top 1 Macro F1 Score: 0.1420
• Top 1 Weighted F1score: 0.1923
• Top 3 Macro F1 Score: 0.2960
• Top3 3 Weighted F1 Score : 0.3853

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: 0.0002
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 2048
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• num_epochs: 5.0

Training results

Training Loss	Epoch	Step	Validation Loss	Top 1 Macro F1 Score	Top 1 Weighted F1score	Top 3 Macro F1 Score	Top3 3 Weighted F1 Score
3.2498	0.14	64	3.0549	0.1275	0.1797	0.2717	0.3636
3.1662	0.27	128	3.0345	0.1311	0.1856	0.2767	0.3674
3.16	0.41	192	3.0158	0.1342	0.1886	0.2792	0.3724
3.1154	0.54	256	3.0034	0.1343	0.1880	0.2844	0.3770
3.1257	0.68	320	2.9848	0.1332	0.1876	0.2820	0.3783
3.1227	0.81	384	2.9721	0.1339	0.1887	0.2844	0.3806
3.1058	0.95	448	2.9737	0.1369	0.1916	0.2871	0.3815
3.0486	1.09	512	3.0066	0.1353	0.1873	0.2888	0.3794
3.0286	1.22	576	3.0036	0.1359	0.1884	0.2874	0.3808
3.0181	1.36	640	2.9924	0.1369	0.1894	0.2888	0.3823
3.0251	1.49	704	2.9899	0.1387	0.1916	0.2878	0.3814
3.0482	1.63	768	2.9787	0.1377	0.1900	0.2903	0.3828
3.0565	1.77	832	2.9746	0.1397	0.1928	0.2912	0.3847
3.0563	1.9	896	2.9710	0.1390	0.1936	0.2887	0.3839
2.9363	2.04	960	3.0022	0.1413	0.1952	0.2902	0.3822
2.951	2.17	1024	3.0072	0.1381	0.1908	0.2902	0.3822
2.9839	2.31	1088	3.0027	0.1413	0.1947	0.2890	0.3811
2.9714	2.45	1152	2.9972	0.1411	0.1945	0.2905	0.3829
2.9959	2.58	1216	2.9788	0.1387	0.1905	0.2906	0.3845
2.9932	2.72	1280	2.9875	0.1412	0.1954	0.2905	0.3840
2.9952	2.85	1344	2.9728	0.1403	0.1933	0.2935	0.3862
3.0112	2.99	1408	2.9707	0.1399	0.1921	0.2958	0.3879
2.9005	3.13	1472	3.0190	0.1398	0.1923	0.2912	0.3825
2.9153	3.26	1536	3.0224	0.1400	0.1935	0.2923	0.3814
2.9284	3.4	1600	3.0087	0.1387	0.1917	0.2895	0.3819
2.922	3.53	1664	3.0130	0.1397	0.1918	0.2927	0.3838
2.947	3.67	1728	3.0023	0.1402	0.1923	0.2916	0.3834
2.9538	3.8	1792	2.9948	0.1414	0.1937	0.2944	0.3845
2.9478	3.94	1856	2.9876	0.1420	0.1952	0.2926	0.3852
2.8445	4.08	1920	3.0409	0.1403	0.1912	0.2961	0.3843
2.8498	4.21	1984	3.0378	0.1412	0.1933	0.2942	0.3813
2.8519	4.35	2048	3.0405	0.1422	0.1942	0.2952	0.3825
2.8715	4.48	2112	3.0341	0.1412	0.1941	0.2922	0.3815
2.892	4.62	2176	3.0262	0.1405	0.1936	0.2912	0.3813
2.8952	4.75	2240	3.0241	0.1428	0.1953	0.2944	0.3840
2.9135	4.89	2304	3.0134	0.1420	0.1923	0.2960	0.3853

Framework versions

• Transformers 4.20.1
• Pytorch 1.12.1+cu102
• Datasets 2.0.0
• Tokenizers 0.11.0