Qwen3-VL-Embedding-8B with MMEB evaluation

Mirror of Qwen/Qwen3-VL-Embedding-8B with the local Nexus MMEB evaluation source, configuration and runtime snapshot, prepared on 2026-10-15. The weights retain their upstream provenance and license. This is not a new fine-tuned model.

Original model documentation and inference examples: UPSTREAM_README.md.
Evaluation setup and runnable commands: evaluation/README.md.
Integrity and source information: package_manifest.json.

The evaluation package includes image, video, moment and visual-document retrieval configurations. Evaluation data must be obtained separately. No new benchmark scores are asserted by this mirror.

Downloads last month: 15

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for woooooz/Qwen3-VL-Embedding-8B

Base model

Qwen/Qwen3-VL-8B-Instruct

Finetuned

Qwen/Qwen3-VL-Embedding-8B

Finetuned

(6)

this model