cstr/qwen3-embed-8b-GGUF

qwen3-embed-8b GGUF

GGUF format of Qwen/Qwen3-Embedding-8B for use with CrispEmbed.

Qwen3 Embedding 8B. Official Alibaba embedding model, 4096-dimensional last-token pooling.

Files

File	Quantization	Size
qwen3-embed-8b-q4_k.gguf	Q4_K	4362 MB
qwen3-embed-8b-q5_k.gguf	Q5_K	5190 MB
qwen3-embed-8b-q8_0.gguf	Q8_0	7674 MB
qwen3-embed-8b.gguf	F32	28873 MB

Quick Start

# Download
huggingface-cli download cstr/qwen3-embed-8b-GGUF qwen3-embed-8b-q4_k.gguf --local-dir .

# Run with CrispEmbed
./crispembed -m qwen3-embed-8b-q4_k.gguf "Hello world"

# Or with auto-download
./crispembed -m qwen3-embed-8b "Hello world"

Model Details

Property	Value
Architecture	Qwen3
Parameters	8B
Embedding Dimension	4096
Layers	36
Pooling	last-token
Tokenizer	GPT-2 BPE
Base Model	Qwen/Qwen3-Embedding-8B

Verification

Verified bit-identical to HuggingFace sentence-transformers (cosine similarity >= 0.999 on test texts).

Usage with CrispEmbed

CrispEmbed is a lightweight C/C++ text embedding inference engine using ggml. No Python runtime, no ONNX. Supports BERT, XLM-R, Qwen3, and Gemma3 architectures.

# Build CrispEmbed
git clone https://github.com/CrispStrobe/CrispEmbed
cd CrispEmbed
cmake -S . -B build && cmake --build build -j

# Encode
./build/crispembed -m qwen3-embed-8b-q4_k.gguf "query text"

# Server mode
./build/crispembed-server -m qwen3-embed-8b-q4_k.gguf --port 8080
curl -X POST http://localhost:8080/v1/embeddings \
    -d '{"input": ["Hello world"], "model": "qwen3-embed-8b"}'

Credits

• Original model: Qwen/Qwen3-Embedding-8B
• Inference engine: CrispEmbed (ggml-based)
• Conversion: convert-decoder-embed-to-gguf.py

Provenance and EU AI Act Art. 53 note

• Upstream model: Qwen/Qwen3-Embedding-8B — published by Qwen.
• Upstream licence: apache-2.0. This repository redistributes under the same terms; it grants no rights the upstream licence does not.
• What was done here: format conversion and/or quantisation only (GGUF/GGML). No training, no fine-tuning, no merging, no distillation, no change to architecture, vocabulary or capability. Only the numeric representation of the upstream weights differs.
• Training data: documented — where it is documented at all — by the upstream provider; see the upstream model card. No training data was used, added or selected by this repository. No training-content summary was found on the upstream model card at the time of writing; that documentation gap is upstream's and is not filled here.
• Provider status: under Regulation (EU) 2024/1689 the upstream authors remain the provider of this model. Converting the serialisation format does not make this repository the provider of a new general-purpose AI model, and no such claim is made. Questions about training content, copyright policy or model capability belong upstream.