<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
IC_ver6J_coco_swin_gpt2_50B_1e
This model is a fine-tuned version of VK246/IC_ver6I_coco_swin_gpt2_50A_1e on the coco dataset. It achieves the following results on the evaluation set:
- Loss: 0.8136
- Cider: 72.7549
- Rouge1: 41.4855
- Rouge2: 16.0212
- Rougel: 37.5584
- Rougelsum: 37.557
- Bleu-1: 42.5255
- Bleu-2: 24.4018
- Bleu-3: 15.3344
- Bleu-4: 10.136
- Gen Len: 11.3063
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 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.5339 | 0.34 | 1000 | 0.8717 | 70.3228 | 40.8259 | 15.4266 | 36.9266 | 36.9254 | 42.2006 | 23.9845 | 14.9776 | 9.8321 | 11.3063 |
0.6269 | 0.68 | 2000 | 0.8136 | 72.7549 | 41.4855 | 16.0212 | 37.5584 | 37.557 | 42.5255 | 24.4018 | 15.3344 | 10.136 | 11.3063 |
Framework versions
- Transformers 4.32.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3