scenario-kd-pre-ner-full-mdeberta_data-univner_en44

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: 63.4449
• Precision: 0.7687
• Recall: 0.7329
• F1: 0.7504
• Accuracy: 0.9792

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: 44
• 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
151.3206	1.2755	500	106.3878	0.1899	0.0932	0.125	0.9466
89.5111	2.5510	1000	81.9211	0.6302	0.6263	0.6282	0.9721
73.496	3.8265	1500	72.8445	0.7147	0.7184	0.7166	0.9779
66.28	5.1020	2000	68.7826	0.7554	0.7226	0.7386	0.9789
62.3735	6.3776	2500	66.0329	0.7339	0.7453	0.7396	0.9786
60.0392	7.6531	3000	64.2510	0.7576	0.7474	0.7525	0.9799
58.3307	8.9286	3500	63.4449	0.7687	0.7329	0.7504	0.9792

Framework versions

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