computer_parts_classifier-model

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.5140
• Accuracy: 0.8069

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.97	9	1.0689	0.5379
1.1042	1.95	18	0.9123	0.6897
0.9605	2.92	27	0.7676	0.7379
0.7855	4.0	37	0.6722	0.7586
0.626	4.97	46	0.5915	0.8069
0.5102	5.95	55	0.5672	0.8138
0.4266	6.92	64	0.5106	0.8483
0.3561	8.0	74	0.5587	0.8138
0.3126	8.97	83	0.5492	0.8069
0.294	9.95	92	0.5589	0.7862
0.2287	10.92	101	0.5579	0.8069
0.2282	12.0	111	0.5193	0.8138
0.2261	12.97	120	0.4383	0.8552
0.2261	13.95	129	0.5205	0.7931
0.1996	14.92	138	0.5037	0.8138
0.1796	16.0	148	0.4986	0.8138
0.1583	16.97	157	0.5583	0.7931
0.1692	17.95	166	0.4743	0.8276
0.1577	18.92	175	0.4867	0.8345
0.1706	19.46	180	0.5140	0.8069

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2