README.md

metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
model-index:
  - name: KcELECTRA-small-v2022-finetuned-in-vehicle
    results: []
language:
  - ko
pipeline_tag: text-classification
widget:
  - text: 오늘 날씨 알려줘.
    example_title: Example 1
  - text: 방탄소년단 노래 찾아줘.
    example_title: Example 2

KcELECTRA-small-v2022-finetuned-in-vehicle

This model is a fine-tuned version of beomi/KcELECTRA-small-v2022 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5573
• Accuracy: 0.8908
• F1: 0.8752

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: 64
• eval_batch_size: 64
• 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	Accuracy	F1
3.3864	1.0	150	3.3222	0.0992	0.0179
3.2362	2.0	300	3.0647	0.2342	0.1422
2.9257	3.0	450	2.7001	0.3467	0.2377
2.5953	4.0	600	2.3780	0.46	0.3409
2.3033	5.0	750	2.1041	0.5025	0.3911
2.0477	6.0	900	1.8675	0.5583	0.4585
1.8309	7.0	1050	1.6711	0.6167	0.5354
1.6455	8.0	1200	1.5126	0.6667	0.5924
1.4883	9.0	1350	1.3677	0.7325	0.6660
1.3473	10.0	1500	1.2435	0.7542	0.6914
1.23	11.0	1650	1.1460	0.7517	0.6911
1.1299	12.0	1800	1.0577	0.7792	0.7251
1.0437	13.0	1950	0.9860	0.7908	0.7397
0.9683	14.0	2100	0.9219	0.8092	0.7625
0.8973	15.0	2250	0.8622	0.8342	0.8028
0.8417	16.0	2400	0.8129	0.8408	0.8122
0.7891	17.0	2550	0.7719	0.85	0.8264
0.743	18.0	2700	0.7323	0.8583	0.8353
0.7037	19.0	2850	0.7005	0.8583	0.8341
0.6715	20.0	3000	0.6696	0.8717	0.8494
0.6396	21.0	3150	0.6469	0.8767	0.8546
0.6117	22.0	3300	0.6285	0.8808	0.8613
0.5897	23.0	3450	0.6137	0.88	0.8618
0.5684	24.0	3600	0.5956	0.88	0.8604
0.5522	25.0	3750	0.5851	0.8808	0.8622
0.5429	26.0	3900	0.5741	0.8867	0.8697
0.53	27.0	4050	0.5672	0.8892	0.8730
0.5234	28.0	4200	0.5616	0.8892	0.8733
0.5184	29.0	4350	0.5580	0.89	0.8738
0.5138	30.0	4500	0.5573	0.8908	0.8752

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2