uisikdag/weed_vit_balanced

weeds_hfclass11

Model is trained on balanced dataset/ 250 image per class/ .8 .1 .1 split/ 224x224 resized

Dataset: https://www.kaggle.com/datasets/vbookshelf/v2-plant-seedlings-dataset

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: 0.3603
• Accuracy: 0.9567

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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.3089	0.99	37	2.0422	0.7133
1.4465	1.99	74	1.2227	0.8767
0.8455	2.99	111	0.8121	0.9067
0.6579	3.99	148	0.6161	0.9267
0.5163	4.99	185	0.5031	0.94
0.4374	5.99	222	0.4078	0.9633
0.3912	6.99	259	0.4134	0.9467
0.358	7.99	296	0.4207	0.9233
0.3509	8.99	333	0.3768	0.95
0.3288	9.99	370	0.3603	0.9567

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu117
• Datasets 2.10.1
• Tokenizers 0.13.2