PassbyGrocer/hreb-msra

robert_bilstm_mega_res-ner-msra-ner-ner-msra-ner

This model is a fine-tuned version of hfl/chinese-roberta-wwm-ext-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0621
• Precision: 0.9538
• Recall: 0.9573
• F1: 0.9555
• Accuracy: 0.9940

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: 64
• eval_batch_size: 64
• seed: 42
• 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: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0239	1.0	725	0.0232	0.9242	0.9344	0.9293	0.9931
0.0139	2.0	1450	0.0254	0.9373	0.9459	0.9416	0.9925
0.006	3.0	2175	0.0294	0.9415	0.9480	0.9448	0.9930
0.0052	4.0	2900	0.0303	0.9389	0.9486	0.9437	0.9937
0.0049	5.0	3625	0.0303	0.9422	0.9498	0.9459	0.9933
0.0034	6.0	4350	0.0353	0.9411	0.9594	0.9502	0.9934
0.0015	7.0	5075	0.0372	0.9404	0.9498	0.9450	0.9927
0.0013	8.0	5800	0.0379	0.9477	0.9492	0.9485	0.9938
0.0006	9.0	6525	0.0405	0.9516	0.9502	0.9509	0.9937
0.0039	10.0	7250	0.0442	0.9420	0.9536	0.9478	0.9931
0.0013	11.0	7975	0.0393	0.9479	0.9528	0.9504	0.9936
0.001	12.0	8700	0.0431	0.9455	0.9513	0.9484	0.9933
0.0011	13.0	9425	0.0431	0.9487	0.9425	0.9455	0.9936
0.0003	14.0	10150	0.0425	0.9392	0.9450	0.9421	0.9933
0.0001	15.0	10875	0.0456	0.9475	0.9515	0.9495	0.9937
0.0011	16.0	11600	0.0446	0.9467	0.9471	0.9469	0.9928
0.0002	17.0	12325	0.0500	0.9532	0.9457	0.9495	0.9933
0.0001	18.0	13050	0.0504	0.9479	0.9490	0.9485	0.9929
0.0002	19.0	13775	0.0455	0.9463	0.9527	0.9495	0.9933
0.0013	20.0	14500	0.0471	0.9487	0.9544	0.9515	0.9933
0.0005	21.0	15225	0.0425	0.9491	0.9584	0.9537	0.9936
0.0009	22.0	15950	0.0503	0.9455	0.9555	0.9505	0.9931
0.0003	23.0	16675	0.0474	0.9530	0.9555	0.9543	0.9938
0.0006	24.0	17400	0.0481	0.9531	0.9538	0.9534	0.9937
0.0013	25.0	18125	0.0502	0.9467	0.9534	0.9500	0.9934
0.0001	26.0	18850	0.0517	0.9461	0.9492	0.9476	0.9933
0.0001	27.0	19575	0.0410	0.9536	0.9530	0.9533	0.9937
0.0011	28.0	20300	0.0453	0.9520	0.9498	0.9509	0.9937
0.0007	29.0	21025	0.0444	0.9479	0.9480	0.9479	0.9935
0.0	30.0	21750	0.0498	0.9529	0.9498	0.9513	0.9937
0.0001	31.0	22475	0.0490	0.9514	0.9496	0.9505	0.9935
0.001	32.0	23200	0.0499	0.9495	0.9486	0.9491	0.9934
0.0001	33.0	23925	0.0451	0.9499	0.9557	0.9528	0.9939
0.0002	34.0	24650	0.0469	0.9486	0.9563	0.9525	0.9937
0.0001	35.0	25375	0.0505	0.9568	0.9496	0.9532	0.9938
0.0003	36.0	26100	0.0491	0.9593	0.9525	0.9559	0.9942
0.0005	37.0	26825	0.0432	0.9551	0.9532	0.9542	0.9939
0.0003	38.0	27550	0.0465	0.9536	0.9486	0.9511	0.9937
0.0019	39.0	28275	0.0491	0.9574	0.9469	0.9521	0.9937
0.0	40.0	29000	0.0470	0.9582	0.9534	0.9558	0.9940
0.0008	41.0	29725	0.0477	0.9505	0.9538	0.9522	0.9937
0.0	42.0	30450	0.0544	0.9500	0.9542	0.9521	0.9937
0.0002	43.0	31175	0.0527	0.9571	0.9492	0.9531	0.9938
0.0005	44.0	31900	0.0510	0.9574	0.9513	0.9543	0.9939
0.0006	45.0	32625	0.0478	0.9527	0.9536	0.9532	0.9938
0.0001	46.0	33350	0.0464	0.9559	0.9517	0.9538	0.9937
0.0001	47.0	34075	0.0478	0.9578	0.9530	0.9554	0.9939
0.0	48.0	34800	0.0507	0.9574	0.9515	0.9544	0.9940
0.0	49.0	35525	0.0534	0.9531	0.9534	0.9532	0.9939
0.0004	50.0	36250	0.0512	0.9541	0.9530	0.9536	0.9941
0.0001	51.0	36975	0.0478	0.9549	0.9532	0.9541	0.9940
0.0001	52.0	37700	0.0446	0.9541	0.9555	0.9548	0.9942
0.0	53.0	38425	0.0522	0.9529	0.9509	0.9519	0.9935
0.0001	54.0	39150	0.0507	0.9552	0.9525	0.9538	0.9937
0.0003	55.0	39875	0.0493	0.9466	0.9484	0.9475	0.9930
0.0	56.0	40600	0.0496	0.9507	0.9496	0.9501	0.9934
0.0	57.0	41325	0.0502	0.9512	0.9559	0.9535	0.9940
0.0	58.0	42050	0.0528	0.9465	0.9525	0.9494	0.9932
0.0	59.0	42775	0.0578	0.9480	0.9503	0.9492	0.9931
0.0	60.0	43500	0.0557	0.9506	0.9486	0.9496	0.9935
0.0	61.0	44225	0.0487	0.9539	0.9521	0.9530	0.9936
0.0	62.0	44950	0.0519	0.9534	0.9536	0.9535	0.9938
0.0	63.0	45675	0.0532	0.9531	0.9554	0.9542	0.9939
0.0	64.0	46400	0.0572	0.9534	0.9527	0.9530	0.9938
0.0001	65.0	47125	0.0563	0.9550	0.9527	0.9538	0.9940
0.0	66.0	47850	0.0550	0.9568	0.9507	0.9538	0.9940
0.0	67.0	48575	0.0585	0.9480	0.9542	0.9511	0.9935
0.0003	68.0	49300	0.0607	0.9501	0.9496	0.9499	0.9936
0.0	69.0	50025	0.0577	0.9529	0.9548	0.9539	0.9939
0.0	70.0	50750	0.0583	0.9541	0.9569	0.9555	0.9941
0.0001	71.0	51475	0.0549	0.9530	0.9486	0.9508	0.9938
0.0	72.0	52200	0.0592	0.9546	0.9509	0.9528	0.9937
0.0	73.0	52925	0.0598	0.9524	0.9502	0.9513	0.9936
0.0	74.0	53650	0.0583	0.9530	0.9517	0.9523	0.9937
0.0	75.0	54375	0.0602	0.9513	0.9513	0.9513	0.9936
0.0	76.0	55100	0.0624	0.9510	0.9527	0.9518	0.9934
0.0	77.0	55825	0.0622	0.9523	0.9527	0.9525	0.9935
0.0	78.0	56550	0.0599	0.9509	0.9536	0.9522	0.9938
0.0	79.0	57275	0.0599	0.9509	0.9550	0.9529	0.9937
0.0	80.0	58000	0.0588	0.9551	0.9536	0.9544	0.9939
0.0	81.0	58725	0.0581	0.9547	0.9561	0.9554	0.9941
0.0	82.0	59450	0.0587	0.9574	0.9567	0.9571	0.9940
0.0	83.0	60175	0.0592	0.9533	0.9582	0.9558	0.9940
0.0	84.0	60900	0.0602	0.9534	0.9569	0.9551	0.9939
0.0	85.0	61625	0.0601	0.9530	0.9554	0.9542	0.9938
0.0	86.0	62350	0.0608	0.9528	0.9561	0.9545	0.9939
0.0	87.0	63075	0.0606	0.9560	0.9538	0.9549	0.9939
0.0	88.0	63800	0.0590	0.9514	0.9575	0.9544	0.9940
0.0	89.0	64525	0.0611	0.9542	0.9577	0.9559	0.9940
0.0002	90.0	65250	0.0617	0.9563	0.9567	0.9565	0.9940
0.0	91.0	65975	0.0611	0.9578	0.9555	0.9566	0.9940
0.0004	92.0	66700	0.0628	0.9510	0.9567	0.9539	0.9939
0.0	93.0	67425	0.0634	0.9523	0.9561	0.9542	0.9939
0.0	94.0	68150	0.0629	0.9534	0.9571	0.9552	0.9940
0.0	95.0	68875	0.0627	0.9523	0.9565	0.9544	0.9940
0.0	96.0	69600	0.0627	0.9528	0.9565	0.9547	0.9940
0.0	97.0	70325	0.0625	0.9536	0.9565	0.9550	0.9940
0.0	98.0	71050	0.0620	0.9558	0.9561	0.9559	0.9941
0.0	99.0	71775	0.0620	0.9543	0.9573	0.9558	0.9940
0.0	100.0	72500	0.0621	0.9538	0.9573	0.9555	0.9940

Framework versions

• Transformers 4.47.1
• Pytorch 2.3.0+cu118
• Datasets 3.2.0
• Tokenizers 0.21.0