Scikit-LLM

Scikit-LLM is an open-source Python library that brings LLM text classification, summarization and translation to scikit-learn.

Scikit-LLM wraps large language models as scikit-learn estimators, so text tasks use the usual fit and predict calls and slot into existing pipelines. It covers zero-shot, few-shot, dynamic few-shot and chain-of-thought text classification, including multi-label classification, plus summarization, translation, text vectorization and an experimental explainable named entity recognition estimator. Estimators are grouped by API format: the GPT family works with OpenAI, Azure OpenAI and local GGUF models through llama.cpp, and a separate family uses Google Vertex AI. Tunable estimators fine-tune the underlying model in the cloud, so no local GPU is needed, and the vectorizer turns text into embeddings for models such as XGBoost. It is maintained by the BeastByte open-source community under the MIT license.

Scikit-LLM at a glance

Best for: Software engineer, Researcher
Runs on: open source
Pricing: Free and open source (MIT); model API usage is billed by the provider
Released: May 2023
Category: AI coding tool
Website: github.com/BeastByteAI/scikit-llm

Key features

Zero-shot and few-shot text classification with LLMs
Dynamic few-shot classification that picks relevant examples per sample
Chain-of-thought and multi-label classifiers
Summarization, translation and explainable named entity recognition
GPTVectorizer to embed text for any scikit-learn model
Tunable estimators that fine-tune models in the cloud
OpenAI, Azure OpenAI, local GGUF and Vertex AI backends

Use it for

Classify support tickets or reviews without labelled training data
Add LLM embeddings to an existing scikit-learn pipeline
Summarize or translate text columns in a dataset
Tag named entities in documents with explanations

Alternatives to Scikit-LLM

AirLLM — Runs massive 70B+ parameter models on small 4GB consumer GPUs.
GLiNER — Enables zero-shot entity extraction without labeled training data on consumer CPUs.
LLM CLI — It enables command-line interaction with remote APIs and local machine models.
llm-d — Deploys production-grade LLM serving on existing hardware clusters.
MLX LM — Runs LLMs and fine-tuning natively on Apple silicon hardware.
vLLM — Maximizes throughput with PagedAttention and continuous batching.
LLM Stats — Independent rankings updated continuously from live API metrics.
LightLLM — It delivers high-speed inference and serving performance for large language models.
MLC LLM — Enables native optimization and deployment of AI models across platforms.
LitGPT — Readable single-file LLM code with tested finetuning recipes
Overview — Learn language models by running hands-on experiments without writing scripts
WebLLM — Runs high-performance language models directly inside web browsers.

All Scikit-LLM alternatives