albert-base-v2-grammar-ner

This model is a fine-tuned version of albert/albert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1134
• Accuracy: 0.9870
• F1 Macro: 0.7941
• F1 Micro: 0.9008
• Precision Macro: 0.8789
• Precision Micro: 0.9569
• Recall Macro: 0.7518
• Recall Micro: 0.8510

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: 24
• eval_batch_size: 24
• 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: 18

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Micro	Precision Macro	Precision Micro	Recall Macro	Recall Micro
0.4297	1.0	93	0.2896	0.9313	0.1318	0.4462	0.1897	0.5163	0.1281	0.3928
0.2521	2.0	186	0.2192	0.9452	0.2315	0.5160	0.3282	0.6752	0.1962	0.4176
0.167	3.0	279	0.1630	0.9662	0.3546	0.7198	0.4142	0.8358	0.3295	0.6321
0.1026	4.0	372	0.1343	0.9733	0.4185	0.7769	0.5241	0.8732	0.3797	0.6998
0.0718	5.0	465	0.1231	0.9738	0.4644	0.7794	0.5584	0.8525	0.4382	0.7178
0.0483	6.0	558	0.1269	0.9778	0.4778	0.8204	0.6262	0.9415	0.4164	0.7269
0.0335	7.0	651	0.1162	0.9804	0.6028	0.8416	0.6985	0.8834	0.5846	0.8036
0.0233	8.0	744	0.1203	0.9813	0.5736	0.8475	0.7429	0.9496	0.4988	0.7652
0.0171	9.0	837	0.1052	0.9836	0.6502	0.8671	0.7023	0.8964	0.6490	0.8397
0.01	10.0	930	0.1125	0.9805	0.6681	0.8477	0.6854	0.8535	0.6875	0.8420
0.0084	11.0	1023	0.1058	0.9862	0.7195	0.8894	0.8004	0.9287	0.6870	0.8533
0.0051	12.0	1116	0.1092	0.9870	0.8015	0.9015	0.8810	0.95	0.7612	0.8578
0.0031	13.0	1209	0.1131	0.9865	0.8006	0.8983	0.8827	0.9429	0.7592	0.8578
0.0017	14.0	1302	0.1106	0.9873	0.8058	0.9039	0.8748	0.9525	0.7749	0.8600
0.0012	15.0	1395	0.1111	0.9875	0.7985	0.9058	0.8818	0.9596	0.7576	0.8578
0.0009	16.0	1488	0.1128	0.9870	0.7941	0.9008	0.8789	0.9569	0.7518	0.8510
0.0008	17.0	1581	0.1133	0.9870	0.7941	0.9008	0.8789	0.9569	0.7518	0.8510
0.0008	18.0	1674	0.1134	0.9870	0.7941	0.9008	0.8789	0.9569	0.7518	0.8510

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3