metadata

library_name: transformers
license: mit
base_model: microsoft/git-large-r-coco
tags:
  - generated_from_trainer
datasets:
  - imagefolder
model-index:
  - name: git-large-r-coco-IDB_ADv1_COCOv6-r
    results: []

git-large-r-coco-IDB_ADv1_COCOv6-r

This model is a fine-tuned version of microsoft/git-large-r-coco on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 3.9193
Meteor Score: {'meteor': 0.3883164385938583}

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: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 64
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 5
num_epochs: 80
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Meteor Score
92.0222	3.3333	5	10.6210	{'meteor': 0.05527200117295446}
80.4987	6.6667	10	9.3234	{'meteor': 0.054478429471574795}
70.872	10.0	15	7.9867	{'meteor': 0.06444810159110426}
61.8948	13.3333	20	7.1539	{'meteor': 0.10729657965271497}
55.8843	16.6667	25	6.5323	{'meteor': 0.1313795489011057}
51.3796	20.0	30	6.0059	{'meteor': 0.1667987247524706}
47.2696	23.3333	35	5.5433	{'meteor': 0.1780717354914205}
43.8752	26.6667	40	5.1460	{'meteor': 0.29042241595126117}
40.8073	30.0	45	4.8067	{'meteor': 0.34065022984874654}
38.3594	33.3333	50	4.5292	{'meteor': 0.37185069963423434}
36.2548	36.6667	55	4.3103	{'meteor': 0.3764524310032932}
34.7935	40.0	60	4.1477	{'meteor': 0.38111013533600185}
33.6518	43.3333	65	4.0357	{'meteor': 0.3851995459070333}
32.8234	46.6667	70	3.9670	{'meteor': 0.38872719984995613}
32.4724	50.0	75	3.9318	{'meteor': 0.3883164385938583}
32.2541	53.3333	80	3.9193	{'meteor': 0.3883164385938583}

Framework versions

Transformers 4.46.1
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.20.2