|
--- |
|
license: apache-2.0 |
|
language: |
|
- ru |
|
- en |
|
- zh |
|
- as |
|
- pl |
|
- ii |
|
- es |
|
metrics: |
|
- fastai |
|
- code_eval |
|
pipeline_tag: zero-shot-classification |
|
library_name: transformers |
|
--- |
|
# FELGUK Upscaler Model |
|
|
|
 |
|
|
|
## Overview |
|
|
|
The **FELGUK Upscaler** is a state-of-the-art image upscaling model designed to enhance the resolution of images while preserving and even improving their quality. This model is particularly useful for tasks such as super-resolution, image restoration, and enhancing low-resolution images. |
|
|
|
This repository contains the pre-trained weights and code to use the FELGUK Upscaler model via the Hugging Face `transformers` library. |
|
|
|
## Model Details |
|
|
|
| Attribute | Description | |
|
|-------------------|-----------------------------------------------------------------------------| |
|
| **Model Name** | FELGUK Upscaler | |
|
| **Architecture** | Based on advanced deep learning techniques for image super-resolution. | |
|
| **Input** | Low-resolution images (e.g., 128x128, 256x256) | |
|
| **Output** | High-resolution images (e.g., 512x512, 1024x1024) | |
|
| **Training Data** | Trained on a diverse dataset of high-quality images. | |
|
| **License** | MIT | |
|
|
|
## Results |
|
|
|
### Original Image |
|
<img src="http://post-images.org/photo-page.php?photo=0wYasIUi" alt="Original Image"> |
|
|
|
### Resulting Image |
|
<img src="http://post-images.org/photo-page.php?photo=piNJm97R" alt="Resulting Image"> |
|
|
|
## Installation |
|
|
|
To use the FELGUK Upscaler model, you need to install the `transformers` library from Hugging Face. You can do this using pip: |
|
|
|
```bash |
|
pip install transformers |
|
``` |
|
## Usage by transformers |
|
```bash |
|
from transformers import FelgukUpscaler, FelgukUpscalerProcessor |
|
from PIL import Image |
|
import requests |
|
|
|
# Load the model and processor |
|
model = FelgukUpscaler.from_pretrained("your-username/felguk-upscaler") |
|
processor = FelgukUpscalerProcessor.from_pretrained("your-username/felguk-upscaler") |
|
|
|
# Load an image from a URL or local file |
|
url = "https://example.com/path/to/your/low-res-image.jpg" |
|
image = Image.open(requests.get(url, stream=True).raw) |
|
|
|
# Preprocess the image |
|
inputs = processor(image, return_tensors="pt") |
|
|
|
# Upscale the image |
|
with torch.no_grad(): |
|
outputs = model(**inputs) |
|
|
|
# Get the upscaled image |
|
upscaled_image = outputs.upscaled_image |
|
|
|
# Save or display the upscaled image |
|
upscaled_image.save("upscaled_image.jpg") |
