scenario-kd-pre-ner-full-mdeberta_data-univner_half55

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: 61.2547
• Precision: 0.7817
• Recall: 0.7759
• F1: 0.7788
• Accuracy: 0.9777

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
151.818	0.5828	500	110.4199	0.6013	0.3174	0.4155	0.9370
98.6874	1.1655	1000	90.3650	0.6787	0.6747	0.6767	0.9686
85.3529	1.7483	1500	83.0290	0.7182	0.7231	0.7206	0.9731
77.9004	2.3310	2000	78.6852	0.7196	0.7658	0.7420	0.9748
73.2356	2.9138	2500	75.0504	0.7578	0.7546	0.7562	0.9756
69.3693	3.4965	3000	72.4160	0.7636	0.7354	0.7492	0.9756
66.3787	4.0793	3500	70.1107	0.7674	0.7475	0.7573	0.9759
63.8461	4.6620	4000	68.2856	0.7786	0.7458	0.7618	0.9762
61.7854	5.2448	4500	66.7176	0.7793	0.7544	0.7667	0.9767
60.08	5.8275	5000	65.3188	0.7888	0.7442	0.7659	0.9764
58.4853	6.4103	5500	64.2104	0.7695	0.7710	0.7703	0.9773
57.5973	6.9930	6000	63.2718	0.7735	0.7781	0.7758	0.9775
56.486	7.5758	6500	62.6904	0.7820	0.7642	0.7730	0.9772
55.7943	8.1585	7000	62.0796	0.7687	0.7759	0.7723	0.9770
55.2592	8.7413	7500	61.6029	0.7763	0.7764	0.7763	0.9779
54.787	9.3240	8000	61.4102	0.7831	0.7781	0.7806	0.9781
54.6374	9.9068	8500	61.2547	0.7817	0.7759	0.7788	0.9777

Framework versions

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