metadata

license: mit
base_model: microsoft/mdeberta-v3-base
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: scenario-TCR-NER_data-univner_half
    results: []

scenario-TCR-NER_data-univner_half

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: 0.1170
Precision: 0.8494
Recall: 0.8655
F1: 0.8574
Accuracy: 0.9842

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: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.1168	0.58	500	0.0625	0.8182	0.8512	0.8344	0.9825
0.0433	1.17	1000	0.0594	0.8396	0.8632	0.8512	0.9843
0.0305	1.75	1500	0.0677	0.8296	0.8703	0.8495	0.9836
0.0213	2.33	2000	0.0761	0.8253	0.8833	0.8533	0.9839
0.0185	2.91	2500	0.0738	0.8600	0.8612	0.8606	0.9850
0.012	3.5	3000	0.0784	0.8374	0.8572	0.8471	0.9835
0.0124	4.08	3500	0.0832	0.8363	0.8704	0.8530	0.9843
0.0095	4.66	4000	0.0806	0.8423	0.8713	0.8565	0.9845
0.008	5.24	4500	0.1049	0.8218	0.8625	0.8417	0.9823
0.0071	5.83	5000	0.0879	0.8420	0.8632	0.8525	0.9842
0.0068	6.41	5500	0.0918	0.8507	0.8733	0.8619	0.9846
0.0058	6.99	6000	0.0951	0.8488	0.8667	0.8577	0.9845
0.0047	7.58	6500	0.0991	0.8467	0.8651	0.8558	0.9842
0.0047	8.16	7000	0.1025	0.8603	0.8573	0.8588	0.9845
0.0043	8.74	7500	0.1020	0.8473	0.8678	0.8574	0.9845
0.0031	9.32	8000	0.1085	0.8437	0.8582	0.8509	0.9842
0.0038	9.91	8500	0.1082	0.8602	0.8440	0.8520	0.9839
0.0024	10.49	9000	0.1163	0.8533	0.8544	0.8539	0.9838
0.0038	11.07	9500	0.1139	0.8528	0.8567	0.8548	0.9843
0.0024	11.66	10000	0.1130	0.8619	0.8476	0.8547	0.9841
0.0024	12.24	10500	0.1170	0.8494	0.8655	0.8574	0.9842

Framework versions

Transformers 4.33.3
Pytorch 2.1.1+cu121
Datasets 2.14.5
Tokenizers 0.13.3