IC_ver6K_coco_swin_gpt2_50A_1e

This model is a fine-tuned version of VK246/IC_ver6J_coco_swin_gpt2_50B_1e on the coco dataset. It achieves the following results on the evaluation set:

• Loss: 0.8220
• Cider: 75.3012
• Rouge1: 41.577
• Rouge2: 16.0357
• Rougel: 37.7284
• Rougelsum: 37.7314
• Bleu-1: 42.4688
• Bleu-2: 24.4386
• Bleu-3: 15.3819
• Bleu-4: 10.1667
• 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
0.4803	0.34	1000	0.9220	69.7834	40.3936	14.9568	36.6495	36.6487	41.6496	23.4287	14.4546	9.3691	11.2806
0.5848	0.68	2000	0.8220	75.3012	41.577	16.0357	37.7284	37.7314	42.4688	24.4386	15.3819	10.1667	11.2806

Framework versions

• Transformers 4.32.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3