multimodla-Bitamin

This model is a fine-tuned version of ddobokki/vision-encoder-decoder-vit-gpt2-coco-ko on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0055
• Rouge1: 11.8323
• Rouge2: 6.4292
• Rougel: 11.8967
• Rougelsum: 11.883
• Gen Len: 100.0

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.1504	1.0	2982	0.1027	0.2262	0.2262	0.2262	0.2262	100.0
0.0664	2.0	5964	0.0457	1.7181	1.2785	1.7523	1.7256	100.0
0.0281	3.0	8946	0.0185	5.3843	3.3178	5.4006	5.4032	100.0
0.0131	4.0	11928	0.0081	10.8352	5.9986	10.7776	10.8786	100.0
0.007	5.0	14910	0.0055	11.8323	6.4292	11.8967	11.883	100.0

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1