Sal-base Open-source Model - Trained on an unknown dataset to provide effective support for related tasks

Sal Base

Developed by bolu61

This model was trained from scratch on an unknown dataset and achieved good performance on the evaluation set, providing effective support for related tasks.

Text Classification

Transformers

#High-accuracy text classification #Multi-class ROC optimization #Few-shot learning

Downloads 216

Release Time : 2/11/2025

Model Overview

A general model trained from scratch, performing well on multiple evaluation metrics and suitable for various classification tasks.

Model Features

High accuracy

The overall accuracy on the evaluation set reaches 87.86%, and it performs excellently in multiple categories.

Multi-class support

Supports classification of multiple categories such as Eol, Msg, Cmd, Var, Dff, Pgr, Stk,Itm, Yml, Obj, Tre.

Stable training

The AdamW optimizer and cosine learning rate scheduler are used, and the training process converges stably.

Model Capabilities

Multi-class classification

High-precision prediction

Imbalanced data processing

Use Cases

Text classification

Document classification

Classify different types of documents.

The F1 value for the Pgr category reaches 0.9405.

Message classification

Distinguish different types of messages.

The ROC AUC value for the Msg category reaches 0.9836.

Object recognition

Object classification

Identify and classify different objects.

Both the precision and recall for the Obj category reach 1.0.

🚀 sal-base

This model was trained from scratch on an unknown dataset and achieves high performance on the evaluation set.

🚀 Quick Start

This sal-base model was trained from scratch on an unknown dataset. It has achieved the following results on the evaluation set:

Metric	Value
Loss	0.7435
Accuracy	0.8786
Precision Eol	0.8682
Recall Eol	0.9233
F1 Eol	0.8949
Support Eol	378
Roc Auc Eol	0.9700
Precision Msg	0.7391
Recall Msg	0.5484
F1 Msg	0.6296
Support Msg	31
Roc Auc Msg	0.9836
Precision Cmd	0.2174
Recall Cmd	0.4167
F1 Cmd	0.2857
Support Cmd	12
Roc Auc Cmd	0.9762
Precision Var	0.0
Recall Var	0.0
F1 Var	0.0
Support Var	15
Roc Auc Var	0.9956
Precision Dff	0.0
Recall Dff	0.0
F1 Dff	0.0
Support Dff	4
Roc Auc Dff	0.9990
Precision Pgr	0.8901
Recall Pgr	0.9970
F1 Pgr	0.9405
Support Pgr	333
Roc Auc Pgr	0.9991
Precision Stk	0.9630
Recall Stk	0.65
F1 Stk	0.7761
Support Stk	40
Roc Auc Stk	0.9453
Precision Itm	0.9821
Recall Itm	0.7639
F1 Itm	0.8594
Support Itm	144
Roc Auc Itm	0.9382
Precision Yml	1.0
Recall Yml	0.8537
F1 Yml	0.9211
Support Yml	41
Roc Auc Yml	0.9964
Precision Obj	1.0
Recall Obj	1.0
F1 Obj	1.0
Support Obj	8
Roc Auc Obj	1.0
Precision Tre	1.0
Recall Tre	0.9583
F1 Tre	0.9787
Support Tre	24
Roc Auc Tre	1.0

🔧 Technical Details

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-06
train_batch_size: 10
eval_batch_size: 10
seed: 45242
optimizer: Use OptimizerNames.ADAMW_TORCH 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: 100.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision Eol	Recall Eol	F1 Eol	Support Eol	Roc Auc Eol	Precision Msg	Recall Msg	F1 Msg	Support Msg	Roc Auc Msg	Precision Cmd	Recall Cmd	F1 Cmd	Support Cmd	Roc Auc Cmd	Support Var	Roc Auc Var	Support Dff	Roc Auc Dff	Precision Pgr	Recall Pgr	F1 Pgr	Support Pgr	Roc Auc Pgr	Precision Stk	Recall Stk	F1 Stk	Support Stk	Roc Auc Stk	Precision Itm	Recall Itm	F1 Itm	Support Itm	Roc Auc Itm	Precision Yml	Recall Yml	F1 Yml	Support Yml	Roc Auc Yml	Precision Obj	Recall Obj	F1 Obj	Support Obj	Roc Auc Obj	Precision Tre	Recall Tre	F1 Tre	Support Tre	Roc Auc Tre
2.5311	1.0	76	2.4941	0.0194	0.0	0.0	0.0	378	0.4851	0.0456	0.4516	0.0828	31	0.7149	0.0073	0.1667	0.0140	12	0.4709	15	0.4025	4	0.8850	0.0	0.0	0.0	333	0.6887	0.0	0.0	0.0	40	0.4469	0.1429	0.0208	0.0364	144	0.5840	0.5	0.0244	0.0465	41	0.7767	0.0	0.0	0.0	8	0.1738	0.0	0.0	0.0	24	0.7487
2.1099	2.0	152	2.0052	0.4388	0.4979	0.3095	0.3817	378	0.5048	0.1190	0.1613	0.1370	31	0.8382	0.0609	0.5833	0.1102	12	0.7734	15	0.5176	4	0.8207	0.5325	0.9339	0.6783	333	0.9186	0.0	0.0	0.0	40	0.5659	0.3571	0.0694	0.1163	144	0.7699	0.5	0.0488	0.0889	41	0.8385	0.0	0.0	0.0	8	0.4631	0.0	0.0	0.0	24	0.7796
1.5835	3.0	228	1.3663	0.5777	0.4690	0.8201	0.5967	378	0.7462	0.0	0.0	0.0	31	0.8730	0.0	0.0	0.0	12	0.9100	15	0.8133	4	0.6306	0.7972	0.8498	0.8227	333	0.9584	0.0	0.0	0.0	40	0.6980	0.2	0.0139	0.0260	144	0.8902	0.0	0.0	0.0	41	0.9466	0.0	0.0	0.0	8	0.8749	0.0	0.0	0.0	24	0.8314
1.3092	4.0	304	1.0934	0.6806	0.5640	0.8624	0.6820	378	0.8593	0.0	0.0	0.0	31	0.8850	0.0	0.0	0.0	12	0.9299	15	0.8156	4	0.6399	0.8944	0.9159	0.9050	333	0.9802	0.0	0.0	0.0	40	0.8137	0.6306	0.4861	0.5490	144	0.9450	0.0	0.0	0.0	41	0.9842	0.0	0.0	0.0	8	0.9250	0.0	0.0	0.0	24	0.9508
0.9523	5.0	380	0.9353	0.7262	0.6149	0.8492	0.7133	378	0.8953	0.0	0.0	0.0	31	0.9026	0.0	0.0	0.0	12	0.9492	15	0.8372	4	0.6399	0.8646	0.9970	0.9261	333	0.9892	0.0	0.0	0.0	40	0.8796	0.7661	0.6597	0.7090	144	0.9474	0.0	0.0	0.0	41	0.9895	0.0	0.0	0.0	8	0.9598	0.0	0.0	0.0	24	0.9965
0.8714	6.0	456	0.7562	0.8039	0.7164	0.9153	0.8037	378	0.9491	0.0	0.0	0.0	31	0.9143	0.0	0.0	0.0	12	0.9499	15	0.8298	4	0.7083	0.925	1.0	0.9610	333	0.9978	1.0	0.35	0.5185	40	0.9170	0.7516	0.7986	0.7744	144	0.9590	1.0	0.3171	0.4815	41	0.9919	0.0	0.0	0.0	8	0.9793	1.0	0.2917	0.4516	24	0.9989
0.7294	7.0	532	0.6222	0.8282	0.7596	0.8862	0.8181	378	0.9575	0.0	0.0	0.0	31	0.9330	0.0	0.0	0.0	12	0.9479	15	0.8359	4	0.8302	0.9276	1.0	0.9624	333	0.9986	1.0	0.375	0.5455	40	0.9634	0.7815	0.8194	0.8	144	0.9655	0.7826	0.8780	0.8276	41	0.9924	0.0	0.0	0.0	8	0.9976	0.9412	0.6667	0.7805	24	0.9993
0.5118	8.0	608	0.5242	0.8534	0.7680	0.9021	0.8297	378	0.9634	0.0	0.0	0.0	31	0.9457	0.0	0.0	0.0	12	0.9656	15	0.8963	4	0.9035	0.9302	1.0	0.9638	333	0.9993	1.0	0.375	0.5455	40	0.9755	0.8873	0.875	0.8811	144	0.9720	0.9091	0.9756	0.9412	41	0.9972	0.0	0.0	0.0	8	0.9991	0.9231	1.0	0.96	24	0.9995
0.4041	9.0	684	0.5220	0.8485	0.8020	0.8571	0.8286	378	0.9552	0.8333	0.3226	0.4651	31	0.9577	0.0	0.0	0.0	12	0.9672	15	0.9258	4	0.8845	0.8582	1.0	0.9237	333	0.9990	0.9667	0.725	0.8286	40	0.9734	0.9487	0.7708	0.8506	144	0.9765	0.9091	0.9756	0.9412	41	0.9977	0.75	0.375	0.5	8	0.9993	0.9231	1.0	0.96	24	0.9997
0.4858	10.0	760	0.4709	0.8748	0.7968	0.9127	0.8508	378	0.9622	0.8333	0.3226	0.4651	31	0.9592	0.0	0.0	0.0	12	0.9721	15	0.9589	4	0.8913	0.9328	1.0	0.9652	333	0.9995	1.0	0.55	0.7097	40	0.9741	0.9690	0.8681	0.9158	144	0.9798	0.9286	0.9512	0.9398	41	0.9982	0.75	0.375	0.5	8	0.9994	0.9231	1.0	0.96	24	0.9999
0.252	11.0	836	0.4155	0.8738	0.7949	0.9127	0.8498	378	0.9694	0.8462	0.3548	0.5	31	0.9737	0.375	0.25	0.3	12	0.9741	15	0.9900	4	0.9608	0.9354	1.0	0.9666	333	0.9996	1.0	0.45	0.6207	40	0.9663	0.9615	0.8681	0.9124	144	0.9811	0.9268	0.9268	0.9268	41	0.9984	0.75	0.375	0.5	8	0.9995	0.9231	1.0	0.96	24	1.0000
0.2958	12.0	912	0.4171	0.8699	0.8354	0.9127	0.8723	378	0.9718	0.8125	0.4194	0.5532	31	0.9773	0.6667	0.3333	0.4444	12	0.9763	15	0.9921	4	0.9518	0.8952	1.0	0.9447	333	0.9997	0.9565	0.55	0.6984	40	0.9705	0.8810	0.7708	0.8222	144	0.9817	0.9268	0.9268	0.9268	41	0.9986	0.8571	0.75	0.8	8	0.9998	0.9231	1.0	0.96	24	0.9997
0.3351	13.0	988	0.4281	0.8786	0.8438	0.8862	0.8645	378	0.9676	0.8421	0.5161	0.64	31	0.9798	0.625	0.4167	0.5	12	0.9788	15	0.9867	4	0.9464	0.9074	1.0	0.9514	333	0.9996	1.0	0.575	0.7302	40	0.9758	0.8865	0.8681	0.8772	144	0.9840	0.9048	0.9268	0.9157	41	0.9984	0.8571	0.75	0.8	8	0.9998	0.9231	1.0	0.96	24	1.0
0.1799	14.0	1064	0.3970	0.8806	0.8159	0.9259	0.8674	378	0.9704	0.8462	0.3548	0.5	31	0.9825	0.625	0.4167	0.5	12	0.9790	15	0.9934	4	0.9976	0.9354	1.0	0.9666	333	0.9997	1.0	0.375	0.5455	40	0.9539	0.9143	0.8889	0.9014	144	0.9859	0.9459	0.8537	0.8974	41	0.9961	0.8571	0.75	0.8	8	0.9998	0.96	1.0	0.9796	24	1.0
0.2573	15.0	1140	0.3927	0.8942	0.8589	0.9180	0.8875	378	0.9713	0.8571	0.5806	0.6923	31	0.9833	0.8333	0.4167	0.5556	12	0.9814	15	0.9899	4	0.9922	0.9276	1.0	0.9624	333	0.9997	1.0	0.55	0.7097	40	0.9706	0.8966	0.9028	0.8997	144	0.9851	0.9474	0.8780	0.9114	41	0.9982	1.0	0.75	0.8571	8	0.9998	0.9231	1.0	0.96	24	1.0
0.1706	16.0	1216	0.3886	0.8835	0.8357	0.9286	0.8797	378	0.9739	0.85	0.5484	0.6667	31	0.9848	0.8333	0.4167	0.5556	12	0.9799	15	0.9940	4	0.9963	0.9171	0.9970	0.9554	333	0.9997	1.0	0.475	0.6441	40	0.9615	0.9084	0.8264	0.8655	144	0.9826	0.9487	0.9024	0.925	41	0.9970	1.0	0.75	0.8571	8	0.9998	1.0	1.0	1.0	24	1.0
0.1121	17.0	1292	0.4100	0.8777	0.8540	0.9286	0.8897	378	0.9717	0.8235	0.4516	0.5833	31	0.9849	0.625	0.4167	0.5	12	0.9783	15	0.9931	4	0.9954	0.8904	1.0	0.9420	333	0.9998	1.0	0.625	0.7692	40	0.9721	0.8790	0.7569	0.8134	144	0.9789	1.0	0.8537	0.9211	41	0.9974	1.0	1.0	1.0	8	1.0	1.0	1.0	1.0	24	1.0
0.1414	18.0	1368	0.4740	0.8757	0.8509	0.9206	0.8844	378	0.9646	0.85	0.5484	0.6667	31	0.9854	0.8333	0.4167	0.5556	12	0.9780	15	0.9932	4	0.9966	0.8833	1.0	0.9380	333	0.9996	1.0	0.575	0.7302	40	0.9652	0.8862	0.7569	0.8165	144	0.9758	0.9722	0.8537	0.9091	41	0.9975	1.0	1.0	1.0	8	1.0	1.0	1.0	1.0	24	1.0
0.0935	19.0	1444	0.4503	0.8689	0.8769	0.9233	0.8995	378	0.9722	0.8077	0.6774	0.7368	31	0.9890	0.25	0.4167	0.3125	12	0.9780	15	0.9939	4	0.9983	0.8582	1.0	0.9237	333	0.9997	0.9630	0.65	0.7761	40	0.9524	0.9592	0.6528	0.7769	144	0.9767	0.9737	0.9024	0.9367	41	0.9982	1.0	0.75	0.8571	8	1.0	1.0	1.0	1.0	24	1.0
0.0794	20.0	1520	0.4563	0.8709	0.8886	0.9074	0.8979	378	0.9740	0.875	0.6774	0.7636	31	0.9879	0.7143	0.4167	0.5263	12	0.9795	15	0.9936	4	0.9976	0.8473	1.0	0.9174	333	0.9998	0.9630	0.65	0.7761	40	0.9514	0.8534	0.6875	0.7615	144	0.9779	0.95	0.9268	0.9383	41	0.9996	1.0	1.0	1.0	8	1.0	1.0	1.0	1.0	24	1.0
0.1209	21.0	1596	0.5927	0.8505	0.8014	0.9286	0.8603	378	0.9639	0.85	0.5484	0.6667	31	0.9857	0.8333	0.4167	0.5556	12	0.9813	15	0.9939	4	0.9956	0.8513	0.9970	0.9184	333	0.9997	1.0	0.375	0.5455	40

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご