git-large-r-coco-IDB_ADv1_COCOv2

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: 3.0890
Meteor Score: {'meteor': 0.4786335839834823}

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: 3e-05
train_batch_size: 128
eval_batch_size: 128
seed: 42
gradient_accumulation_steps: 32
total_train_batch_size: 4096
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
num_epochs: 150
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Meteor Score
11.2941	5.0	5	9.7802	{'meteor': 0.047306330587744805}
9.5485	10.0	10	8.8574	{'meteor': 0.07078101725961515}
8.5922	15.0	15	7.9155	{'meteor': 0.08874535718930834}
7.8228	20.0	20	7.3148	{'meteor': 0.12018891275455396}
7.285	25.0	25	6.8648	{'meteor': 0.15707348326414072}
6.8607	30.0	30	6.4858	{'meteor': 0.19388208635297605}
6.4959	35.0	35	6.1520	{'meteor': 0.21120188981070245}
6.1676	40.0	40	5.8462	{'meteor': 0.2716218553024612}
5.8644	45.0	45	5.5574	{'meteor': 0.3167687928552893}
5.578	50.0	50	5.2839	{'meteor': 0.3935488961336244}
5.3073	55.0	55	5.0231	{'meteor': 0.4401242744337809}
5.0514	60.0	60	4.7784	{'meteor': 0.4548021686867608}
4.8117	65.0	65	4.5514	{'meteor': 0.4596569131466085}
4.5893	70.0	70	4.3420	{'meteor': 0.45977845518579624}
4.3842	75.0	75	4.1505	{'meteor': 0.46392172400764786}
4.1966	80.0	80	3.9751	{'meteor': 0.47171195207926336}
4.0273	85.0	85	3.8188	{'meteor': 0.47363666995497383}
3.8755	90.0	90	3.6800	{'meteor': 0.4762423622409033}
3.741	95.0	95	3.5578	{'meteor': 0.4773783003509899}
3.6242	100.0	100	3.4534	{'meteor': 0.477309973037375}
3.5237	105.0	105	3.3645	{'meteor': 0.4793397639012666}
3.4397	110.0	110	3.2905	{'meteor': 0.47738435127434337}
3.3702	115.0	115	3.2312	{'meteor': 0.4783368243538133}
3.3149	120.0	120	3.1840	{'meteor': 0.47570665785547966}
3.2715	125.0	125	3.1485	{'meteor': 0.47646100050859935}
3.2398	130.0	130	3.1234	{'meteor': 0.4749786711607146}
3.2173	135.0	135	3.1059	{'meteor': 0.4790692343269223}
3.2025	140.0	140	3.0958	{'meteor': 0.47859655244912347}
3.1947	145.0	145	3.0907	{'meteor': 0.47856080653742605}
3.1907	150.0	150	3.0890	{'meteor': 0.4786335839834823}

Framework versions

Transformers 4.46.1
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.20.2

加载中…

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 3e-05

train_batch_size: 128

eval_batch_size: 128

seed: 42

gradient_accumulation_steps: 32

total_train_batch_size: 4096

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

num_epochs: 150

mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Meteor Score
11.2941	5.0	5	9.7802	{'meteor': 0.047306330587744805}
9.5485	10.0	10	8.8574	{'meteor': 0.07078101725961515}
8.5922	15.0	15	7.9155	{'meteor': 0.08874535718930834}
7.8228	20.0	20	7.3148	{'meteor': 0.12018891275455396}
7.285	25.0	25	6.8648	{'meteor': 0.15707348326414072}
6.8607	30.0	30	6.4858	{'meteor': 0.19388208635297605}
6.4959	35.0	35	6.1520	{'meteor': 0.21120188981070245}
6.1676	40.0	40	5.8462	{'meteor': 0.2716218553024612}
5.8644	45.0	45	5.5574	{'meteor': 0.3167687928552893}
5.578	50.0	50	5.2839	{'meteor': 0.3935488961336244}
5.3073	55.0	55	5.0231	{'meteor': 0.4401242744337809}
5.0514	60.0	60	4.7784	{'meteor': 0.4548021686867608}
4.8117	65.0	65	4.5514	{'meteor': 0.4596569131466085}
4.5893	70.0	70	4.3420	{'meteor': 0.45977845518579624}
4.3842	75.0	75	4.1505	{'meteor': 0.46392172400764786}
4.1966	80.0	80	3.9751	{'meteor': 0.47171195207926336}
4.0273	85.0	85	3.8188	{'meteor': 0.47363666995497383}
3.8755	90.0	90	3.6800	{'meteor': 0.4762423622409033}
3.741	95.0	95	3.5578	{'meteor': 0.4773783003509899}
3.6242	100.0	100	3.4534	{'meteor': 0.477309973037375}
3.5237	105.0	105	3.3645	{'meteor': 0.4793397639012666}
3.4397	110.0	110	3.2905	{'meteor': 0.47738435127434337}
3.3702	115.0	115	3.2312	{'meteor': 0.4783368243538133}
3.3149	120.0	120	3.1840	{'meteor': 0.47570665785547966}
3.2715	125.0	125	3.1485	{'meteor': 0.47646100050859935}
3.2398	130.0	130	3.1234	{'meteor': 0.4749786711607146}
3.2173	135.0	135	3.1059	{'meteor': 0.4790692343269223}
3.2025	140.0	140	3.0958	{'meteor': 0.47859655244912347}
3.1947	145.0	145	3.0907	{'meteor': 0.47856080653742605}
3.1907	150.0	150	3.0890	{'meteor': 0.4786335839834823}

Framework versions

Transformers 4.46.1

Pytorch 2.2.1+cu121

Datasets 2.18.0

Tokenizers 0.20.2

ooliverz/git-large-r-coco-IDB_ADv1_COCOv2

git-large-r-coco-IDB_ADv1_COCOv2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

git-large-r-coco-IDB_ADv1_COCOv2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions