Candle

Candle provides a minimalist Rust framework for machine learning model development and execution tasks.

github.com/huggingface/candle
Candle homepage screenshot

About Candle

Candle is a minimalist machine learning framework for Rust with a focus on performance, including GPU support, and ease of use. It handles tensor operations on CPU and CUDA GPU devices and provides examples for language models, text-to-image generation, object detection, image segmentation, and audio models including speech recognition and text-to-speech.

Some examples are available as online demos that run entirely in the browser, including YOLO for pose estimation and object recognition, Whisper for speech recognition, LLaMA2 and T5 for text generation, Segment Anything for image segmentation, and BLIP for image captioning.

Key features

  • Tensor operations in Rust with CPU and CUDA device selection
  • Browser demos for object detection, speech recognition, and text generation
  • Command-line examples for LLaMA, Mistral, Falcon, Gemma, and other LLMs
  • Image generation examples with Stable Diffusion and Wuerstchen models
  • Image segmentation examples with Segment Anything, SegFormer, and YOLO
  • Audio examples including Whisper, EnCodec, MetaVoice, and Parler-TTS
Best for: Rust developers who want to develop and run machine learning models with CPU or GPU execution.
Good for
  • Minimalist design focused on performance and ease of use
  • Supports GPU execution including CUDA devices
  • Online demos run entirely in the browser

Candle at a glance

Type
Open-source project
Pricing
Not confirmed. Check the site
Open source
Yes · Source code
Language
English
Last checked
Oct 5, 2026 · Online when checked
Listed
Oct 5, 2026

Candle alternatives

Tools and sites you could use instead, ranked by how closely they match.

All 24 alternatives

Popular in AI Infrastructure & MLOps

Well-known listings from the same topic.

All 341

Candle FAQ

How do I switch from CPU to GPU execution?
After installing Candle with CUDA support, use Device::new_cuda(0) instead of Device::Cpu.
What is needed to start running code?
Install candle-core as described in Installation, then run Rust code such as a matrix multiplication with Tensor and Device.
Is something wrong or out of date?

Tell us and an editor will check it.

Own this website?

Claim the listing and suggest updates yourself. An editor checks every claim first.

Claim this listing