transformers

Looking for HuggingFace Transformers in Rust?

Check out candle-pipelines - simple, intuitive pipelines for local LLM inference in Rust, inspired by Python's Transformers library.

With candle-pipelines:

let pipeline = TextGenerationPipelineBuilder::qwen3(Qwen3Size::Size0_6B).build().await?;

let response = pipeline.completion("What is 1+1?").await?;

println!(response) // 2 + 2 = 4

cargo add candle-pipelines