IC_ver5b_coco_swin_gpt2_01pc_1e

This model is a fine-tuned version of VK246/IC_ver5a_coco_swin_gpt2_05pc_1e on the coco dataset. It achieves the following results on the evaluation set:

• Loss: 1.1266
• Rouge1: 27.4772
• Rouge2: 5.9305
• Rougel: 25.1138
• Rougelsum: 25.1235
• Bleu: 2.437
• 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: 96
• eval_batch_size: 96
• 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
1.2093	0.42	25	1.1552	22.8898	3.6353	20.6781	20.6737	1.1554	11.1124
1.2149	0.85	50	1.1358	26.2857	5.2765	24.0266	24.0308	2.1954	11.1124

Framework versions

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