results

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: 0.0540
• Precision: 0.9528
• Recall: 0.9561
• F1: 0.9545
• Accuracy: 0.9941

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 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0237	1.0	725	0.0217	0.9247	0.9355	0.9301	0.9934
0.0157	2.0	1450	0.0239	0.9358	0.9482	0.9420	0.9934
0.0074	3.0	2175	0.0264	0.9465	0.9507	0.9486	0.9936
0.0032	4.0	2900	0.0288	0.9390	0.9478	0.9434	0.9934
0.0022	5.0	3625	0.0335	0.9409	0.9440	0.9425	0.9932
0.0051	6.0	4350	0.0328	0.9500	0.9513	0.9507	0.9938
0.002	7.0	5075	0.0372	0.9436	0.9471	0.9453	0.9929
0.0007	8.0	5800	0.0447	0.9394	0.9486	0.9440	0.9925
0.0011	9.0	6525	0.0403	0.9488	0.9530	0.9509	0.9936
0.0009	10.0	7250	0.0425	0.9477	0.9596	0.9536	0.9938
0.001	11.0	7975	0.0420	0.9438	0.9436	0.9437	0.9927
0.0005	12.0	8700	0.0495	0.9414	0.9525	0.9469	0.9920
0.0007	13.0	9425	0.0424	0.9485	0.9602	0.9543	0.9938
0.0004	14.0	10150	0.0399	0.9422	0.9536	0.9479	0.9937
0.0002	15.0	10875	0.0360	0.9469	0.9511	0.9490	0.9942
0.0003	16.0	11600	0.0395	0.9468	0.9521	0.9494	0.9936
0.0004	17.0	12325	0.0420	0.9495	0.9544	0.9519	0.9943
0.0011	18.0	13050	0.0400	0.9462	0.9515	0.9489	0.9939
0.0006	19.0	13775	0.0410	0.9473	0.9554	0.9513	0.9940
0.0001	20.0	14500	0.0492	0.9416	0.9525	0.9470	0.9933
0.0002	21.0	15225	0.0426	0.9501	0.9555	0.9528	0.9941
0.0006	22.0	15950	0.0450	0.9459	0.9430	0.9445	0.9934
0.0001	23.0	16675	0.0584	0.9415	0.9450	0.9432	0.9931
0.0	24.0	17400	0.0465	0.9497	0.9557	0.9527	0.9939
0.0	25.0	18125	0.0513	0.9534	0.9569	0.9551	0.9940
0.0001	26.0	18850	0.0500	0.9442	0.9579	0.9510	0.9933
0.0	27.0	19575	0.0468	0.9506	0.9592	0.9549	0.9939
0.0007	28.0	20300	0.0441	0.9513	0.9509	0.9511	0.9936
0.0	29.0	21025	0.0493	0.9494	0.9525	0.9509	0.9938
0.0004	30.0	21750	0.0490	0.9502	0.9548	0.9525	0.9937
0.0	31.0	22475	0.0516	0.9475	0.9509	0.9492	0.9937
0.0	32.0	23200	0.0552	0.9440	0.9530	0.9485	0.9933
0.0	33.0	23925	0.0548	0.9422	0.9534	0.9478	0.9932
0.0002	34.0	24650	0.0523	0.9476	0.9536	0.9506	0.9936
0.0	35.0	25375	0.0461	0.9534	0.9565	0.9549	0.9941
0.0	36.0	26100	0.0482	0.9459	0.9548	0.9503	0.9939
0.0001	37.0	26825	0.0520	0.9522	0.9552	0.9537	0.9939
0.0	38.0	27550	0.0532	0.9525	0.9542	0.9534	0.9937
0.0	39.0	28275	0.0515	0.9502	0.9542	0.9522	0.9939
0.0	40.0	29000	0.0518	0.9540	0.9550	0.9545	0.9941
0.0	41.0	29725	0.0529	0.9529	0.9546	0.9538	0.9939
0.0001	42.0	30450	0.0531	0.9534	0.9565	0.9549	0.9940
0.0	43.0	31175	0.0542	0.9535	0.9557	0.9546	0.9941
0.0	44.0	31900	0.0538	0.9507	0.9548	0.9528	0.9940
0.0	45.0	32625	0.0568	0.9482	0.9538	0.9510	0.9938
0.0001	46.0	33350	0.0548	0.9546	0.9544	0.9545	0.9941
0.0	47.0	34075	0.0545	0.9515	0.9561	0.9538	0.9941
0.0	48.0	34800	0.0538	0.9519	0.9563	0.9541	0.9942
0.0	49.0	35525	0.0540	0.9525	0.9561	0.9543	0.9941
0.0	50.0	36250	0.0540	0.9528	0.9561	0.9545	0.9941

Framework versions

• Transformers 4.46.2
• Pytorch 2.3.0+cu118
• Datasets 3.1.0
• Tokenizers 0.20.3