GIT_inf_w_caption_blur_ep5_eval

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

• Loss: 0.0688
• Rouge1: 11.5483
• Rouge2: 6.9038
• Rougel: 10.6731
• Rougelsum: 10.6966
• Gen Len: 217.74

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.0775	1.0	1586	0.0774	11.837	5.6006	10.9944	11.0165	218.44
0.0658	2.0	3172	0.0726	9.7028	5.0964	9.0714	9.0844	218.44
0.0541	3.0	4758	0.0693	11.4449	6.3978	10.5899	10.6179	218.44
0.0432	4.0	6344	0.0682	11.1405	6.5221	10.3109	10.3318	218.39
0.0342	5.0	7930	0.0688	11.5483	6.9038	10.6731	10.6966	217.74

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.1+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2