maskara-tiny-v2

This model is a fine-tuned version of somukandula/maskara-tiny on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1264
• Precision: 0.0627
• Recall: 0.2523
• F1: 0.1004

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: 8e-06
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
1.2386	0.0936	1000	1.4658	0.0466	0.1926	0.0751
1.1062	0.1873	2000	1.2917	0.0576	0.2296	0.0921
1.0024	0.2809	3000	1.2243	0.0605	0.2386	0.0966
1.0004	0.3745	4000	1.1891	0.0600	0.2425	0.0961
0.962	0.4682	5000	1.1587	0.0641	0.2515	0.1022
0.9539	0.5618	6000	1.1307	0.0627	0.2444	0.0998
0.9417	0.6554	7000	1.1264	0.0627	0.2523	0.1004

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1