convnextv2-base-22k-224-finetuned-tekno24

This model is a fine-tuned version of facebook/convnextv2-base-22k-224 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9120
• Accuracy: 0.6138
• F1: 0.5996
• Precision: 0.5969
• Recall: 0.6138

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.3179	0.9968	78	1.2415	0.4207	0.3979	0.4642	0.4207
1.1998	1.9936	156	1.0769	0.5103	0.4525	0.5309	0.5103
1.168	2.9904	234	1.0573	0.5494	0.5033	0.5605	0.5494
1.1107	4.0	313	0.9924	0.5540	0.5163	0.5257	0.5540
1.1062	4.9968	391	1.0018	0.5747	0.5507	0.5660	0.5747
1.0331	5.9936	469	0.9901	0.5931	0.5768	0.6202	0.5931
1.0409	6.9904	547	0.9634	0.5747	0.5723	0.5722	0.5747
1.0176	8.0	626	0.9504	0.5931	0.5834	0.5814	0.5931
0.995	8.9968	704	0.9584	0.5908	0.5854	0.5853	0.5908
0.9937	9.9936	782	0.9339	0.6023	0.5934	0.5894	0.6023
0.9387	10.9904	860	0.9120	0.6138	0.5996	0.5969	0.6138
0.9324	11.9617	936	0.9135	0.5954	0.5879	0.5865	0.5954

Framework versions

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