metadata

library_name: transformers
license: mit
base_model: microsoft/deberta-v3-small
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: doc-topic-model_eval-01_train-02
    results: []

doc-topic-model_eval-01_train-02

This model is a fine-tuned version of microsoft/deberta-v3-small on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.0379
Accuracy: 0.9877
F1: 0.6213
Precision: 0.7252
Recall: 0.5434

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: 4
eval_batch_size: 256
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 100
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.0944	0.4931	1000	0.0900	0.9814	0.0	0.0	0.0
0.0769	0.9862	2000	0.0688	0.9814	0.0015	0.9091	0.0008
0.0607	1.4793	3000	0.0561	0.9822	0.1054	0.8204	0.0563
0.0535	1.9724	4000	0.0501	0.9845	0.3683	0.7561	0.2434
0.0466	2.4655	5000	0.0451	0.9857	0.4853	0.7323	0.3629
0.0441	2.9586	6000	0.0423	0.9862	0.5089	0.7590	0.3827
0.0391	3.4517	7000	0.0406	0.9866	0.5538	0.7285	0.4467
0.0372	3.9448	8000	0.0395	0.9869	0.5537	0.7576	0.4362
0.0336	4.4379	9000	0.0387	0.9871	0.5704	0.7494	0.4604
0.0337	4.9310	10000	0.0381	0.9872	0.5865	0.7368	0.4871
0.0297	5.4241	11000	0.0374	0.9874	0.6051	0.7282	0.5175
0.0296	5.9172	12000	0.0383	0.9872	0.5796	0.7475	0.4732
0.0263	6.4103	13000	0.0381	0.9873	0.6096	0.7110	0.5335
0.0272	6.9034	14000	0.0380	0.9874	0.6193	0.7078	0.5505
0.0234	7.3964	15000	0.0379	0.9876	0.6178	0.7265	0.5373
0.0243	7.8895	16000	0.0379	0.9877	0.6213	0.7252	0.5434

Framework versions

Transformers 4.44.2
Pytorch 2.4.1+cu121
Datasets 2.21.0
Tokenizers 0.19.1