IC_ver6I_coco_swin_gpt2_50A_1e

This model is a fine-tuned version of VK246/IC_ver6H_coco_swin_gpt2_50B_1e on the coco dataset. It achieves the following results on the evaluation set:

• Loss: 0.8003
• Cider: 36.4847
• Rouge1: 41.9392
• Rouge2: 16.4156
• Rougel: 38.0808
• Rougelsum: 38.0721
• Bleu-1: 42.8624
• Bleu-2: 24.8647
• Bleu-3: 15.7144
• Bleu-4: 10.4434
• 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
0.5622	0.34	1000	0.8598	16.5035	41.0303	15.4795	37.2917	37.2896	41.7661	23.7724	14.7804	9.5941	11.2806
0.639	0.68	2000	0.8003	36.4847	41.9392	16.4156	38.0808	38.0721	42.8624	24.8647	15.7144	10.4434	11.2806

Framework versions

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