pytorch-fid
Package for calculating Frechet Inception Distance (FID) using PyTorch
Decision gist · record as of 2026-08-14
Yes, if you need FID evaluation for generative models in PyTorch. The package is stable, permissively licensed, and has low install friction. However, maintenance is dormant (last release January 2023); verify current compatibility with your PyTorch version first. Results differ slightly from the original implementation, so check whether exact reproducibility matters for your use case.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires PyTorch and torchvision installed; GPU acceleration requires CUDA-capable hardware.
- Low friction install with standard dependencies.
- Dormant maintenance status (last release January 2023) means no active development, but the package is stable.
License · maintenance · safety
permissive license (permissive) — Licensed under Apache License 2.0 (permissive), allowing commercial and private use with minimal restrictions.
last release 2023-01-05 (1317 days) · last repo commit 2024-07-03 · 3,851 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 153,171 downloads/mo, #10,886 on PyPI
Alternatives
Verify before relying
pip install pytorch-fid
python -m pytorch_fid path/to/dataset1 path/to/dataset2
# Or with GPU acceleration:
python -m pytorch_fid --device cuda:0 path/to/dataset1 path/to/dataset2- Whether results remain comparable to official TensorFlow implementation given reported interpolation differences
- Current compatibility with recent PyTorch and torchvision versions beyond requires_python >=3.5
What it is and what it does
pytorch-fid is a PyTorch implementation of the Fréchet Inception Distance metric, originally developed to evaluate generative model quality by measuring statistical distance between real and generated image distributions. It computes FID by fitting Gaussians to feature representations extracted from the Inception network and calculating the Fréchet distance between them.
The package provides a command-line interface to compute FID scores between two image folders, with options to use different Inception layers (64, 192, 768, or 2048 dimensions) and GPU acceleration. It can also pre-compute and cache dataset statistics as .npz files for repeated comparisons. The weights and model match the official implementation, though results may differ slightly due to image interpolation and backend differences.
Use it for
- Evaluate generative adversarial network (GAN) sample quality during training or model comparison
- Measure similarity between real and synthetic image datasets for validation
- Pre-compute and cache statistics from a reference dataset to benchmark multiple models against it
- Compare image generation models using a standardized metric that correlates with human visual quality judgment
- Use lower-dimensional Inception features when comparing datasets with fewer than 2048 images
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you need FID evaluation for generative models in PyTorch.
The package is stable, permissively licensed, and has low install friction. However, maintenance is dormant (last release January 2023); verify current compatibility with your PyTorch version first. Results differ slightly from the original implementation, so check whether exact reproducibility matters for your use case.
Install
pytorch-fid on PyPI
Before you install
Low friction install with standard dependencies. Dormant maintenance status (last release January 2023) means no active development, but the package is stable.
Requires PyTorch and torchvision installed; GPU acceleration requires CUDA-capable hardware.
License in practice
Licensed under Apache License 2.0 (permissive), allowing commercial and private use with minimal restrictions.
Quickstart
pip install pytorch-fid
python -m pytorch_fid path/to/dataset1 path/to/dataset2
# Or with GPU acceleration:
python -m pytorch_fid --device cuda:0 path/to/dataset1 path/to/dataset2Verify before relying
- Whether results remain comparable to official TensorFlow implementation given reported interpolation differences
- Current compatibility with recent PyTorch and torchvision versions beyond requires_python >=3.5
Package facts
| License | permissive license permissive |
| Python support | Supports the current Python release >=3.5 |
| Install friction | Low. Pure-Python wheel |
| Runtime dependencies | 5 packagesnumpypillowscipytorchtorchvision |
| Maintenance | Dormant 1,317 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 153,171 / month, #10,886 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | License :: OSI Approved :: Apache Software LicenseProgramming Language :: Python :: 3 |
Evidence: pytorch_fid-0.3.0-py3-none-any.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “image quality evaluation metric”
- pytorch-fidComputes Fréchet Inception Distance (FID), a metric for measuring…
- bert-scoreBERTScore computes automatic evaluation metrics for text generation…
- lpipsComputes perceptual similarity between image pairs using deep neural…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also clean-fid · piq · lpips · tomesd · facenet-pytorch · pyiqa · pretrainedmodels · effdet · pytorchcv