Downloads · 30 days
0
vit-cvt-cifar10-experiments
vivekdhayaal/vit-cvt-cifar10-experiments
vit-cvt-cifar10-experiments is a machine learning model from vivekdhayaal. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
This repository contains the experimental results, training curves, and attention map visualizations for a custom implementation of a Vision Transformer (ViT) and Convolutional Vision Transformer (CvT) trained from sc…
Access
Public
Updated Apr 19, 2026
Repo size
4.8 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.png3.7 MB · 100%
At a glance
- Access
- Public
- Created
- Apr 18, 2026
- Updated
- Apr 19, 2026
- SHA
- 7b374c99
Model card
From the Hugging Face model README
Vision Transformer & Convolutional ViT on CIFAR-10
This repository contains the experimental results, training curves, and attention map visualizations for a custom implementation of a Vision Transformer (ViT) and Convolutional Vision Transformer (CvT) trained from scratch on CIFAR-10.
Directory Structure
code/: Custom PyTorch implementations of Scaled Dot-Product Attention, Transformer Blocks, and standard/convolutional patch embeddings.training_curves/: Loss and accuracy plots covering patch size variations, hyperparameter tuning, data augmentations, and positional embedding ablations.visualizations/: Attention maps across heads, Attention Rollout flow, and Positional Embedding cosine similarity matrices.
- Languages
- en
- Created
- Apr 18, 2026
- Updated
- Apr 19, 2026