convnextv2-base-22k-224-finetuned-tekno23-24-mean-std

This model is a fine-tuned version of facebook/convnextv2-base-22k-224 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2419
• Accuracy: 0.9024
• F1: 0.9023
• Precision: 0.9055
• Recall: 0.9024

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.7083	0.9995	275	0.5788	0.7620	0.7615	0.7730	0.7620
0.513	1.9991	550	0.4341	0.8154	0.8133	0.8372	0.8154
0.462	2.9986	825	0.3308	0.8604	0.8579	0.8606	0.8604
0.4301	3.9982	1100	0.3219	0.8663	0.8637	0.8695	0.8663
0.4029	4.9977	1375	0.3042	0.8758	0.8755	0.8779	0.8758
0.3795	5.9973	1650	0.2419	0.9024	0.9023	0.9055	0.9024
0.3444	6.9968	1925	0.2493	0.8962	0.8960	0.8982	0.8962
0.337	8.0	2201	0.2311	0.9018	0.9012	0.9030	0.9018
0.2933	8.9995	2476	0.2405	0.8988	0.8983	0.9000	0.8988
0.3416	9.9955	2750	0.2334	0.9018	0.9015	0.9033	0.9018

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1