IC_ver6G_coco_swin_gpt2_50A_1e

This model is a fine-tuned version of VK246/IC_ver6F_coco_swin_gpt2_50B_1e on the coco dataset. It achieves the following results on the evaluation set:

• Loss: 0.7892
• Cider: 15.3553
• Rouge1: 41.9548
• Rouge2: 16.3636
• Rougel: 38.1268
• Rougelsum: 38.1269
• Bleu-1: 42.7082
• Bleu-2: 24.7427
• Bleu-3: 15.632
• Bleu-4: 10.351
• Gen Len: 11.2806

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	Cider	Rouge1	Rouge2	Rougel	Rougelsum	Bleu-1	Bleu-2	Bleu-3	Bleu-4	Gen Len
0.6462	0.34	1000	0.8195	16.4275	41.2411	15.6826	37.4271	37.4332	42.1612	24.0307	15.0144	9.8204	11.2806
0.6923	0.68	2000	0.7892	15.3553	41.9548	16.3636	38.1268	38.1269	42.7082	24.7427	15.632	10.351	11.2806

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3