PromptSite provides prompt version management for tracking, experimentation, and debugging LLM prompts.
Open source
24 alternativesOneComp
OneCompression is a Python package for compressing large language models.
About OneComp
OneComp is an open-source Python package for compressing large language models with post-training quantization. It runs from the command line as onecomp or from Python with Runner.auto_run, detecting available GPU VRAM, assigning per-layer bit widths, quantizing, evaluating, and saving the result.
It includes quantization methods such as QEP, LPCD, AutoBit, JointQ, MDBF, and block-wise distillation, along with LoRA fine-tuning, rotation preprocessing, vLLM plugins for serving, and GGUF export for llama.cpp/Ollama with Hugging Face Hub upload. GGUF export supports Llama and Qwen2 architectures, and rotation preprocessing supports Llama and Qwen3 architectures.
Key features
- Automatic VRAM-based per-layer bit-width assignment via AutoBit
- Quantization error propagation to subsequent layers with QEP
- Serve quantized models with vLLM plugins for DBF, Mixed-GPTQ
- Convert models to GGUF v3 format for llama.cpp and Ollama
- LoRA fine-tuning of quantized models for accuracy recovery
- Block-wise distillation against FP16 teacher at block granularity
Best for: Researchers and developers who need to quantize LLMs to fit available GPU memory.
Good for
- Single-command quantization via CLI or Python Runner
- Multiple quantization methods including QEP, LPCD, JointQ, and MDBF
- Integrations for vLLM serving, GGUF export, and Hugging Face Hub upload
Watch out for
- GGUF export supports only Llama and Qwen2 architectures
- Rotation preprocessing supports only Llama and Qwen3 architectures
OneComp 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
OneComp alternatives
Tools and sites you could use instead, ranked by how closely they match.
DeepEval is an LLM evaluation framework, according to its repository description.
Open source
24 alternatives
LiteLLM provides a way to call more than 100 language models through an OpenAI-compatible interface.
This repository contains utilities intended for use with Llama models.
Ollama runs language models locally and provides tools to work with model files and APIs.
Open source
24 alternatives
OpenLLMetry provides OpenTelemetry-based observability for generative AI and large language model applications.
Open source
24 alternativesPopular in AI Infrastructure & MLOps
Well-known listings from the same topic.
Guild AIguild.aiFlow AIflow-ai.comsystemprompt.iosystemprompt.ioKili Technologykili-technology.comAPI docsdocs.x.aiEyePop.aieyepop.ai
OneComp FAQ
How do I start automatic quantization?
Run `onecomp meta-llama/Llama-2-7b-hf` from the command line, or call `Runner.auto_run(model_id="meta-llama/Llama-2-7b-hf")` from Python.
How does AutoBit choose bit widths?
It estimates the target bit width from available VRAM and assigns per-layer bit widths to minimize quantization error under the memory budget.
How can quantized models be used for inference?
Serve them with vLLM using built-in plugins for DBF and Mixed-GPTQ methods, or convert them to GGUF v3 format for llama.cpp and Ollama.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.