bert-sentiment-classifier

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0820
• Accuracy: 0.9736
• F1: 0.9736
• Precision: 0.9736
• Recall: 0.9736
• F1 Negative: 0.9736
• F1 Neutral: 0.9736
• F1 Positive: 0.0

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: 2.0000000000000003e-06
• train_batch_size: 128
• eval_batch_size: 256
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 2
• total_train_batch_size: 512
• total_eval_batch_size: 512
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Accuracy	F1	F1 Negative	F1 Neutral	F1 Positive	Validation Loss	Precision	Recall
0.1503	0.0711	1000	0.9526	0.9526	0.9529	0.9524	0.0	0.1399	0.9527	0.9526
0.123	0.1422	2000	0.9613	0.9613	0.9612	0.9614	0.0	0.1169	0.9613	0.9613
0.1085	0.2133	3000	0.9623	0.9623	0.9619	0.9627	0.0	0.1103	0.9625	0.9623
0.1071	0.2845	4000	0.9658	0.9658	0.9656	0.9659	0.0	0.0997	0.9658	0.9658
0.1048	0.3556	5000	0.9669	0.9669	0.9671	0.9668	0.0	0.0973	0.9670	0.9669
0.0952	0.4267	6000	0.9674	0.9674	0.9676	0.9671	0.0	0.1002	0.9674	0.9674
0.098	0.4978	7000	0.9689	0.9689	0.9689	0.9689	0.0	0.0952	0.9689	0.9689
0.0967	0.5689	8000	0.9689	0.9689	0.9689	0.9690	0.0	0.0930	0.9689	0.9689
0.0936	0.6400	9000	0.9693	0.9693	0.9691	0.9695	0.0	0.0926	0.9694	0.9693
0.0904	0.7111	10000	0.9691	0.9691	0.9689	0.9694	0.0	0.0946	0.9693	0.9691
0.0943	0.7823	11000	0.9700	0.9700	0.9698	0.9701	0.0	0.0880	0.9700	0.9700
0.0921	0.8534	12000	0.9703	0.9703	0.9701	0.9704	0.0	0.0867	0.9703	0.9703
0.0867	0.9245	13000	0.9704	0.9704	0.9702	0.9706	0.0	0.0878	0.9704	0.9704
0.0863	0.9956	14000	0.9707	0.9707	0.9706	0.9708	0.0	0.0871	0.9707	0.9707
0.0798	1.0667	15000	0.9709	0.9709	0.9710	0.9709	0.0	0.0883	0.9710	0.9709
0.0772	1.1378	16000	0.9711	0.9711	0.9710	0.9712	0.0	0.0871	0.9712	0.9711
0.0759	1.2089	17000	0.9719	0.9719	0.9719	0.9719	0.0	0.0884	0.9719	0.9719
0.0767	1.2800	18000	0.9717	0.9717	0.9715	0.9718	0.0	0.0857	0.9717	0.9717
0.0791	1.3512	19000	0.9718	0.9718	0.9717	0.9719	0.0	0.0870	0.9718	0.9718
0.0766	1.4223	20000	0.9722	0.9722	0.9722	0.9721	0.0	0.0827	0.9722	0.9722
0.0808	1.4934	21000	0.9725	0.9725	0.9725	0.9725	0.0	0.0829	0.9725	0.9725
0.0784	1.5645	22000	0.9726	0.9726	0.9726	0.9725	0.0	0.0824	0.9726	0.9726
0.0814	1.6356	23000	0.9727	0.9727	0.9727	0.9727	0.0	0.0811	0.9727	0.9727
0.0789	1.7067	24000	0.9727	0.9727	0.9727	0.9727	0.0	0.0825	0.9727	0.9727
0.0762	1.7778	25000	0.9734	0.9734	0.9734	0.9734	0.0	0.0806	0.9734	0.9734
0.0766	1.8490	26000	0.9732	0.9732	0.9731	0.9732	0.0	0.0813	0.9732	0.9732
0.0764	1.9201	27000	0.9728	0.9728	0.9727	0.9730	0.0	0.0825	0.9729	0.9728
0.0737	1.9912	28000	0.9732	0.9732	0.9733	0.9730	0.0	0.0818	0.9732	0.9732
0.0644	2.0623	29000	0.9733	0.9733	0.9732	0.9733	0.0	0.0835	0.9733	0.9733
0.0678	2.1334	30000	0.9732	0.9732	0.9732	0.9732	0.0	0.0841	0.9732	0.9732
0.0653	2.2045	31000	0.9734	0.9734	0.9734	0.9734	0.0	0.0842	0.9734	0.9734
0.0639	2.2756	32000	0.9734	0.9734	0.9734	0.9735	0.0	0.0827	0.9734	0.9734
0.0617	2.3468	33000	0.9734	0.9734	0.9733	0.9734	0.0	0.0835	0.9734	0.9734
0.0645	2.4179	34000	0.9734	0.9734	0.9735	0.9734	0.0	0.0824	0.9734	0.9734
0.0623	2.4890	35000	0.9733	0.9733	0.9733	0.9734	0.0	0.0827	0.9734	0.9733
0.0602	2.5601	36000	0.9734	0.9734	0.9734	0.9734	0.0	0.0833	0.9734	0.9734
0.0625	2.6312	37000	0.9734	0.9734	0.9733	0.9734	0.0	0.0830	0.9734	0.9734
0.0649	2.7023	38000	0.9734	0.9734	0.9734	0.9735	0.0	0.0825	0.9734	0.9734
0.0574	2.7734	39000	0.9735	0.9735	0.9735	0.9735	0.0	0.0828	0.9735	0.9735
0.0632	2.8445	40000	0.9736	0.9736	0.9736	0.9736	0.0	0.0821	0.9736	0.9736
0.0643	2.9157	41000	0.9736	0.9736	0.9736	0.9736	0.0	0.0820	0.9736	0.9736
0.0605	2.9868	42000	0.9736	0.9736	0.9736	0.9736	0.0	0.0820	0.9736	0.9736
0.0648	3.0579	43000	0.0896	0.9723	0.9723	0.9723	0.9723	0.9724	0.9722	0.0
0.0748	3.1290	44000	0.0896	0.9716	0.9716	0.9717	0.9716	0.9714	0.9718	0.0
0.067	3.2001	45000	0.0862	0.9722	0.9722	0.9722	0.9722	0.9723	0.9721	0.0
0.0708	3.2712	46000	0.0861	0.9726	0.9726	0.9726	0.9726	0.9726	0.9725	0.0
0.0678	3.3423	47000	0.0857	0.9724	0.9724	0.9724	0.9724	0.9723	0.9724	0.0
0.071	3.4135	48000	0.0877	0.9722	0.9722	0.9722	0.9722	0.9722	0.9721	0.0
0.0717	3.4846	49000	0.0846	0.9719	0.9719	0.9720	0.9719	0.9718	0.9721	0.0
0.0731	3.5557	50000	0.0836	0.9722	0.9722	0.9722	0.9722	0.9722	0.9722	0.0
0.073	3.6268	51000	0.0825	0.9726	0.9726	0.9726	0.9726	0.9727	0.9725	0.0
0.0726	3.6979	52000	0.0823	0.9728	0.9728	0.9728	0.9728	0.9727	0.9728	0.0
0.0686	3.7690	53000	0.0826	0.9728	0.9728	0.9728	0.9728	0.9728	0.9729	0.0
0.068	3.8401	54000	0.0824	0.9728	0.9728	0.9728	0.9728	0.9728	0.9728	0.0
0.0713	3.9113	55000	0.0835	0.9728	0.9728	0.9729	0.9728	0.9729	0.9728	0.0
0.0706	3.9824	56000	0.0827	0.9725	0.9725	0.9725	0.9725	0.9724	0.9726	0.0
0.0601	4.0535	57000	0.0844	0.9726	0.9726	0.9726	0.9726	0.9726	0.9725	0.0
0.0589	4.1246	58000	0.0879	0.9724	0.9724	0.9724	0.9724	0.9724	0.9724	0.0
0.0618	4.1957	59000	0.0868	0.9725	0.9725	0.9725	0.9725	0.9725	0.9725	0.0
0.0609	4.2668	60000	0.0876	0.9724	0.9724	0.9725	0.9724	0.9725	0.9723	0.0
0.0605	4.3379	61000	0.0934	0.9716	0.9716	0.9717	0.9716	0.9714	0.9718	0.0
0.0588	4.4090	62000	0.0929	0.9728	0.9728	0.9728	0.9728	0.9727	0.9728	0.0
0.0613	4.4802	63000	0.0875	0.9723	0.9723	0.9723	0.9723	0.9722	0.9724	0.0
0.0591	4.5513	64000	0.0888	0.9726	0.9726	0.9726	0.9726	0.9726	0.9726	0.0
0.0618	4.6224	65000	0.0856	0.9724	0.9724	0.9725	0.9724	0.9725	0.9724	0.0
0.0564	4.6935	66000	0.0884	0.9727	0.9727	0.9727	0.9727	0.9727	0.9727	0.0
0.0629	4.7646	67000	0.0854	0.9729	0.9729	0.9729	0.9729	0.9729	0.9729	0.0
0.0601	4.8357	68000	0.0881	0.9729	0.9729	0.9729	0.9729	0.9729	0.9729	0.0
0.0619	4.9068	69000	0.0857	0.9726	0.9726	0.9726	0.9726	0.9726	0.9726	0.0
0.0591	4.9780	70000	0.0857	0.9726	0.9726	0.9726	0.9726	0.9725	0.9727	0.0
0.0499	5.0491	71000	0.0895	0.9725	0.9725	0.9725	0.9725	0.9725	0.9726	0.0
0.0526	5.1202	72000	0.0912	0.9724	0.9724	0.9724	0.9724	0.9723	0.9724	0.0
0.0543	5.1913	73000	0.0943	0.9727	0.9727	0.9727	0.9727	0.9726	0.9727	0.0
0.0526	5.2624	74000	0.0920	0.9723	0.9723	0.9723	0.9723	0.9723	0.9724	0.0
0.0576	5.3335	75000	0.0901	0.9721	0.9721	0.9721	0.9721	0.9720	0.9722	0.0
0.0518	5.4046	76000	0.0951	0.9718	0.9718	0.9718	0.9718	0.9717	0.9719	0.0
0.0533	5.4758	77000	0.0898	0.9722	0.9722	0.9722	0.9722	0.9722	0.9722	0.0
0.0485	5.5469	78000	0.0941	0.9722	0.9722	0.9722	0.9722	0.9721	0.9722	0.0
0.052	5.6180	79000	0.0909	0.9725	0.9725	0.9725	0.9725	0.9724	0.9725	0.0
0.0505	5.6891	80000	0.0957	0.9721	0.9721	0.9722	0.9721	0.9720	0.9722	0.0
0.0525	5.7602	81000	0.0934	0.9723	0.9723	0.9723	0.9723	0.9723	0.9723	0.0
0.0491	5.8313	82000	0.0921	0.9724	0.9724	0.9724	0.9724	0.9723	0.9724	0.0
0.0538	5.9024	83000	0.0920	0.9725	0.9725	0.9725	0.9725	0.9724	0.9725	0.0
0.0516	5.9735	84000	0.0917	0.9725	0.9725	0.9725	0.9725	0.9724	0.9725	0.0
0.041	6.0447	85000	0.1008	0.9721	0.9721	0.9721	0.9721	0.9719	0.9722	0.0
0.0424	6.1158	86000	0.1065	0.9719	0.9719	0.9719	0.9719	0.9718	0.9720	0.0
0.0442	6.1869	87000	0.0982	0.9722	0.9722	0.9722	0.9722	0.9722	0.9723	0.0
0.0435	6.2580	88000	0.1031	0.9721	0.9721	0.9721	0.9721	0.9720	0.9722	0.0
0.0421	6.3291	89000	0.1003	0.9719	0.9719	0.9719	0.9719	0.9719	0.9719	0.0
0.0408	6.4002	90000	0.1028	0.9720	0.9720	0.9720	0.9720	0.9719	0.9720	0.0
0.0441	6.4713	91000	0.0981	0.9718	0.9718	0.9718	0.9718	0.9717	0.9719	0.0
0.0447	6.5425	92000	0.0952	0.9723	0.9723	0.9723	0.9723	0.9723	0.9723	0.0
0.0461	6.6136	93000	0.0949	0.9720	0.9720	0.9720	0.9720	0.9720	0.9720	0.0
0.0439	6.6847	94000	0.0988	0.9722	0.9722	0.9722	0.9722	0.9722	0.9723	0.0
0.0443	6.7558	95000	0.0988	0.9720	0.9720	0.9720	0.9720	0.9719	0.9720	0.0
0.0395	6.8269	96000	0.1013	0.9723	0.9723	0.9723	0.9723	0.9724	0.9722	0.0
0.0414	6.8980	97000	0.1010	0.9724	0.9724	0.9724	0.9724	0.9724	0.9724	0.0
0.0469	6.9691	98000	0.0998	0.9722	0.9722	0.9722	0.9722	0.9722	0.9723	0.0
0.0329	7.0403	99000	0.1126	0.9721	0.9721	0.9721	0.9721	0.9721	0.9721	0.0
0.038	7.1114	100000	0.1076	0.9714	0.9714	0.9714	0.9714	0.9713	0.9715	0.0
0.0374	7.1825	101000	0.1045	0.9722	0.9722	0.9722	0.9722	0.9722	0.9722	0.0
0.0377	7.2536	102000	0.1080	0.9718	0.6479	0.6479	0.6479	0.9717	0.0	0.9719
0.0393	7.3247	103000	0.1036	0.9722	0.9722	0.9722	0.9722	0.9722	0.9722	0.0
0.0393	7.3958	104000	0.1045	0.9722	0.9722	0.9722	0.9722	0.9722	0.9722	0.0
0.0382	7.4669	105000	0.1081	0.9718	0.9718	0.9718	0.9718	0.9717	0.9719	0.0
0.0374	7.5380	106000	0.1010	0.9719	0.9719	0.9719	0.9719	0.9719	0.9719	0.0
0.0355	7.6092	107000	0.1092	0.9717	0.9717	0.9718	0.9717	0.9716	0.9719	0.0
0.035	7.6803	108000	0.1101	0.9721	0.9721	0.9721	0.9721	0.9720	0.9721	0.0
0.0344	7.7514	109000	0.1089	0.9722	0.9722	0.9722	0.9722	0.9721	0.9722	0.0
0.0348	7.8225	110000	0.1095	0.9721	0.9721	0.9721	0.9721	0.9721	0.9721	0.0
0.0387	7.8936	111000	0.1062	0.9717	0.9717	0.9717	0.9717	0.9716	0.9718	0.0
0.0352	7.9647	112000	0.1076	0.9722	0.9722	0.9722	0.9722	0.9722	0.9722	0.0
0.0331	8.0358	113000	0.1149	0.9719	0.9719	0.9719	0.9719	0.9718	0.9719	0.0
0.0362	8.1070	114000	0.1122	0.9718	0.9718	0.9718	0.9718	0.9718	0.9718	0.0
0.0357	8.1781	115000	0.1107	0.9714	0.9714	0.9714	0.9714	0.9713	0.9715	0.0
0.0328	8.2492	116000	0.1136	0.9719	0.9719	0.9719	0.9719	0.9719	0.9719	0.0
0.0327	8.3203	117000	0.1178	0.9717	0.9717	0.9717	0.9717	0.9717	0.9718	0.0
0.0315	8.3914	118000	0.1154	0.9715	0.6477	0.6477	0.6477	0.9714	0.0	0.9716
0.031	8.4625	119000	0.1137	0.9717	0.9717	0.9717	0.9717	0.9717	0.9718	0.0
0.0314	8.5336	120000	0.1128	0.9718	0.9718	0.9718	0.9718	0.9718	0.9718	0.0
0.0352	8.6048	121000	0.1111	0.9717	0.9717	0.9717	0.9717	0.9716	0.9717	0.0
0.0356	8.6759	122000	0.1112	0.9718	0.9718	0.9718	0.9718	0.9718	0.9719	0.0
0.0345	8.7470	123000	0.1135	0.9718	0.9718	0.9718	0.9718	0.9718	0.9719	0.0
0.0351	8.8181	124000	0.1144	0.9718	0.9718	0.9718	0.9718	0.9717	0.9718	0.0
0.0311	8.8892	125000	0.1149	0.9719	0.9719	0.9719	0.9719	0.9719	0.9720	0.0
0.0353	8.9603	126000	0.1118	0.9718	0.9718	0.9718	0.9718	0.9718	0.9718	0.0
0.0331	9.0314	127000	0.1162	0.9717	0.9717	0.9717	0.9717	0.9717	0.9718	0.0
0.0282	9.1025	128000	0.1181	0.9716	0.6478	0.6478	0.6478	0.9716	0.0	0.9717
0.0307	9.1737	129000	0.1176	0.9717	0.6478	0.6478	0.6478	0.9716	0.0	0.9717
0.0297	9.2448	130000	0.1192	0.9716	0.6477	0.6478	0.6477	0.9716	0.0	0.9717
0.0334	9.3159	131000	0.1174	0.9717	0.6478	0.6478	0.6478	0.9716	0.0	0.9717
0.0295	9.3870	132000	0.1178	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718
0.0321	9.4581	133000	0.1168	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718
0.0323	9.5292	134000	0.1169	0.9717	0.6478	0.6478	0.6478	0.9716	0.0	0.9717
0.0295	9.6003	135000	0.1174	0.9717	0.6478	0.6478	0.6478	0.9716	0.0	0.9718
0.0327	9.6715	136000	0.1179	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718
0.0323	9.7426	137000	0.1176	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718
0.0275	9.8137	138000	0.1177	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718
0.0339	9.8848	139000	0.1176	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718
0.0325	9.9559	140000	0.1175	0.9717	0.6478	0.6478	0.6478	0.9717	0.0	0.9718

Framework versions

• Transformers 4.57.1
• Pytorch 2.9.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1