IC_ver4a_coco_swin_gpt2_1pc_1e

This model is a fine-tuned version of on the coco dataset. It achieves the following results on the evaluation set:

• Loss: 1.0910
• Rouge1: 29.7994
• Rouge2: 7.3652
• Rougel: 27.9817
• Rougelsum: 28.041
• Bleu: 2.5557
• Gen Len: 11.1124

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

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bleu	Gen Len
1.7489	0.14	50	1.2748	22.1874	2.634	21.043	21.059	0.0	11.1124
1.3067	0.28	100	1.1917	25.4818	4.8904	23.8981	23.981	2.9734	11.1124
1.2822	0.42	150	1.1982	24.3457	4.1568	22.512	22.528	0.0	11.1124
1.2584	0.56	200	1.1801	23.7424	3.9363	22.0734	22.0981	0.0	11.1124
1.2148	0.71	250	1.1412	25.134	5.0362	23.6681	23.7093	2.0303	11.1124
1.1806	0.85	300	1.1048	28.8206	7.4959	27.0534	27.1111	2.965	11.1124
1.2013	0.99	350	1.0911	29.6942	7.3408	27.8612	27.901	2.5222	11.1124

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3