Downloads · 30 days
0
vit_base_patch16_pc_parts_classifier
shaanzeeeee/vit_base_patch16_pc_parts_classifier
vit_base_patch16_pc_parts_classifier is a image classification model from shaanzeeeee. Use it when you need a label for an image. It is set up for fastai. The card lists the license as mit.
Vision Transformer image classifier for 11 PC component and cable-management classes.
Access
Public
Updated Apr 16, 2026
Repo size
690 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pkl345 MB · 50%
At a glance
- Task
- Image Classification
- Library
- fastai
- License
- mit
- Access
- Public
- Created
- Apr 16, 2026
- Updated
- Apr 16, 2026
- SHA
- 6e6df1df
Model card
From the Hugging Face model README
vit_base_patch16_pc_parts_classifier
Vision Transformer image classifier for 11 PC component and cable-management classes.
Model Details
- Architecture: ViT-Base Patch16 224 (
vit_base_patch16_224) - Framework: FastAI + timm + PyTorch
- Input size: 224x224 RGB
- Classes: 11
- Epochs: 15
- Batch size: 16
- Test accuracy: 0.7389
- Training date: 2026-04-17
Labels
- AIO_Liquid_Cooler
- Air_Cooler
- Bad_Cable_Management
- CPU
- Good_Cable_Management
- Graphics_Card
- M2_NVMe_Drive
- Motherboard
- PC_Case
- Power_Supply
- RAM_Stick
Files
best_model_export.pkl: FastAI export for direct inference.best_model_state_dict.pth: PyTorch state dict.best_model_metadata.json: Training and class metadata.
Inference (FastAI)
from fastai.learner import load_learner
from pathlib import Path
learn = load_learner("best_model_export.pkl")
pred_class, pred_idx, probs = learn.predict(Path("sample.jpg"))
print(pred_class)
print({learn.dls.vocab[i]: float(probs[i]) for i in range(len(probs))})
Live Demo
A live inference demo is available on Hugging Face Spaces:
- Task
- Image Classification
- Library
- fastai
- License
- mit
- Languages
- en
- Created
- Apr 16, 2026
- Updated
- Apr 16, 2026