cstr/qwen3-vl-2b-crispembed-gguf

Qwen3-VL-2B — CrispEmbed GGUF

GGUF conversions of Qwen/Qwen3-VL-2B-Instruct for use with CrispEmbed.

Models

File	Quant	Size	Description
qwen3-vl-2b-q4_k.gguf	Q4_K	1.5 GB	Good quality/size balance
qwen3-vl-2b-q8_0.gguf	Q8_0	2.2 GB	Best quality

Features

• DeepStack vision fusion: intermediate vision-encoder features injected into LLM decoder layers
• Fused flash attention: uses ggml_flash_attn_ext for efficient inference
• Backend KV cache: decode stays on GPU (Metal/CUDA), no per-token CPU transfer
• Interleaved mRoPE: improved position encoding vs Qwen2.5-VL
• QK RMSNorm: per-head query/key normalization

Usage

0 ""

0 ""

0 ""

1 "/usr/include/stdc-predef.h" 1 3 4

0 "" 2

1 ""

Converted with from CrispEmbed.

Provenance and EU AI Act Art. 53 note

• Upstream model: Qwen/Qwen3-VL-2B-Instruct — 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). 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.