README.md

metadata

license: mit
tags:
  - generated_from_keras_callback
model-index:
  - name: madatnlp/gamza-bart-for-kormath
    results: []

madatnlp/gamza-bart-for-kormath

This model is a fine-tuned version of gogamza/kobart-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Train Loss: 0.1418
• Validation Loss: 0.3009
• Epoch: 29

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:

• optimizer: {'name': 'Adam', 'learning_rate': 1e-04, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
• training_precision: float32

Training results

Train Loss	Validation Loss	Epoch
4.4155	1.9300	0
1.4995	1.0293	1
1.0445	0.8365	2
0.8775	0.7569	3
0.8198	0.7778	4
0.7619	0.7430	5
0.7324	0.7259	6
0.7234	0.7214	7
0.6697	0.6819	8
0.6599	0.6673	9
0.6387	0.6433	10
0.6227	0.6651	11
0.6017	0.6128	12
0.5820	0.6430	13
0.5229	0.5611	14
0.4617	0.4675	15
0.4071	0.4463	16
0.3495	0.4213	17
0.3202	0.4103	18
0.2875	0.4477	19
0.2528	0.3244	20
0.2331	0.4037	21
0.2117	0.3041	22
0.1943	0.3069	23
0.1805	0.3385	24
0.2267	0.3347	25
0.2049	0.2993	26
0.1800	0.3792	27
0.1583	0.2905	28
0.1418	0.3009	29

Framework versions

• Transformers 4.18.0
• TensorFlow 2.8.0
• Datasets 2.1.0
• Tokenizers 0.12.1