deberta-v3-base_fine_tuned_food_ner

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

• Loss: 0.4164
• Precision: 0.9268
• Recall: 0.9446
• F1: 0.9356
• Accuracy: 0.9197

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	40	0.8425	0.8323	0.8323	0.8323	0.8073
No log	2.0	80	0.5533	0.8703	0.8941	0.8820	0.8731
No log	3.0	120	0.4855	0.8771	0.9109	0.8937	0.8797
No log	4.0	160	0.4238	0.8949	0.9222	0.9083	0.8964
No log	5.0	200	0.4176	0.9048	0.9302	0.9173	0.9008
No log	6.0	240	0.4127	0.9065	0.9342	0.9202	0.9004
No log	7.0	280	0.4409	0.9294	0.9302	0.9298	0.9043
No log	8.0	320	0.3971	0.9129	0.9334	0.9230	0.9061
No log	9.0	360	0.3941	0.9112	0.9390	0.9249	0.9061
No log	10.0	400	0.4069	0.9233	0.9366	0.9299	0.9148
No log	11.0	440	0.4039	0.9213	0.9390	0.9300	0.9162
No log	12.0	480	0.4000	0.9126	0.9470	0.9295	0.9113
0.3799	13.0	520	0.4126	0.9323	0.9390	0.9356	0.9179
0.3799	14.0	560	0.4076	0.9272	0.9398	0.9334	0.9140
0.3799	15.0	600	0.4129	0.9317	0.9414	0.9365	0.9188
0.3799	16.0	640	0.4000	0.9239	0.9446	0.9341	0.9162
0.3799	17.0	680	0.4098	0.9267	0.9438	0.9352	0.9179
0.3799	18.0	720	0.4110	0.9232	0.9454	0.9342	0.9188
0.3799	19.0	760	0.4202	0.9275	0.9446	0.9360	0.9183
0.3799	20.0	800	0.4164	0.9268	0.9446	0.9356	0.9197

Framework versions

• Transformers 4.21.0
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1