Scholarlink

In this research paper, PyTorch’s Neural Style Transfer (NST) feature is utilized to establish a seamless way for combining the content of one image with the appearance of another image. Users can submit two different images: content image and style image from the user interface. After uploading, the NST process is applied on those input images with the help of GAN model, and the output is finally produced. The user can upload their own images. With Streamlit, an intuitive platform is provided to the users to explore various possibilities by trying out different inputs. The generated output is then displayed in the UI where users can access it. This application is a place for artists, designers, and enthusiasts to test and implement their creativity. The FID score of GAN model using PyTorch is 25.0 and 30.0 using TensorFlow. So PyTorch is more suitable for this application.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Neural Style Transfer Using PyTorch

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Syed Ashraf,
  • Ganesh B. Regulwar,
  • Uppu Karthikeya,
  • Pentala Srinith Reddy,
  • K. Dhanalaxmi,
  • Srithin Chillamcharla

摘要

In this research paper, PyTorch’s Neural Style Transfer (NST) feature is utilized to establish a seamless way for combining the content of one image with the appearance of another image. Users can submit two different images: content image and style image from the user interface. After uploading, the NST process is applied on those input images with the help of GAN model, and the output is finally produced. The user can upload their own images. With Streamlit, an intuitive platform is provided to the users to explore various possibilities by trying out different inputs. The generated output is then displayed in the UI where users can access it. This application is a place for artists, designers, and enthusiasts to test and implement their creativity. The FID score of GAN model using PyTorch is 25.0 and 30.0 using TensorFlow. So PyTorch is more suitable for this application.