scenario-kd-pre-ner-full-mdeberta_data-univner_full55

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 46.7361
• Precision: 0.8159
• Recall: 0.8270
• F1: 0.8214
• Accuracy: 0.9820

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 32
• seed: 55
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
152.0575	0.2911	500	111.6819	0.2800	0.0814	0.1261	0.9323
101.0937	0.5822	1000	93.6121	0.7508	0.5637	0.6439	0.9650
89.9589	0.8732	1500	86.3988	0.7477	0.7279	0.7377	0.9744
83.7555	1.1643	2000	81.5901	0.8025	0.7064	0.7514	0.9753
78.7764	1.4554	2500	77.3559	0.7740	0.7852	0.7795	0.9787
75.052	1.7465	3000	74.0454	0.7888	0.7958	0.7923	0.9795
71.8558	2.0375	3500	71.2511	0.8026	0.7785	0.7904	0.9795
68.6774	2.3286	4000	68.6385	0.7945	0.7976	0.7960	0.9799
66.1585	2.6197	4500	66.3761	0.8064	0.7911	0.7987	0.9800
64.1799	2.9108	5000	64.3746	0.8086	0.8038	0.8062	0.9805
61.921	3.2019	5500	62.3790	0.8003	0.8113	0.8058	0.9806
60.0217	3.4929	6000	60.7116	0.8101	0.8124	0.8113	0.9810
58.3481	3.7840	6500	59.3713	0.8109	0.8173	0.8141	0.9812
57.0197	4.0751	7000	57.8634	0.8097	0.8147	0.8122	0.9810
55.4761	4.3662	7500	56.6438	0.8111	0.8142	0.8126	0.9810
54.2509	4.6573	8000	55.5117	0.8185	0.8150	0.8167	0.9812
53.179	4.9483	8500	54.4697	0.8090	0.8166	0.8128	0.9814
52.0052	5.2394	9000	53.5329	0.8134	0.8224	0.8178	0.9812
51.025	5.5305	9500	52.6787	0.8205	0.8179	0.8192	0.9816
50.2537	5.8216	10000	51.7982	0.8124	0.8227	0.8175	0.9815
49.562	6.1126	10500	51.1222	0.8153	0.8227	0.8190	0.9814
48.6986	6.4037	11000	50.3991	0.8198	0.8321	0.8259	0.9820
48.0743	6.6948	11500	49.9026	0.8108	0.8273	0.8190	0.9814
47.4624	6.9859	12000	49.2992	0.8227	0.8257	0.8242	0.9822
46.8068	7.2770	12500	48.7934	0.8200	0.8339	0.8269	0.9819
46.4058	7.5680	13000	48.3805	0.8159	0.8272	0.8215	0.9818
46.0154	7.8591	13500	48.0425	0.8188	0.8189	0.8189	0.9814
45.6997	8.1502	14000	47.7079	0.8146	0.8240	0.8193	0.9814
45.3236	8.4413	14500	47.3817	0.8239	0.8241	0.8240	0.9821
45.0217	8.7324	15000	47.1403	0.8237	0.8299	0.8268	0.9820
44.7989	9.0234	15500	46.9718	0.8174	0.8228	0.8201	0.9818
44.6325	9.3145	16000	46.8718	0.8192	0.8237	0.8214	0.9816
44.5013	9.6056	16500	46.7607	0.8217	0.8263	0.8240	0.9820
44.4881	9.8967	17000	46.7361	0.8159	0.8270	0.8214	0.9820

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1