---
base_model:
- ShutterMuse/ShutterMuse
library_name: transformers
tags:
- text-generation-inference
- llama-cpp
- qwen3-vl
- vision-language
- multimodal
- photography-guidance
language:
- en
pipeline_tag: image-text-to-text
license: other
---

# **ShutterMuse-GGUF**

> **ShutterMuse** is a multimodal large language model for capture-time photography guidance, built on Qwen3-VL-8B-Instruct and described in the paper "ShutterMuse: Capture-Time Photography Guidance with MLLMs" (arXiv:2606.25763). As the title indicates, it is meant to give guidance while a photo is being taken. The model card itself is minimal. It does not describe the training data, the training method, the types of guidance produced, any benchmark results, or the license, so those details are best taken from the paper. The checkpoint is a bfloat16 sharded safetensors release using the `Qwen3VLForConditionalGeneration` architecture and `Qwen3VLProcessor`, and it loads through standard Transformers. **ShutterMuse** on HF — [hf.co/ShutterMuse](https://huggingface.co/ShutterMuse/ShutterMuse).

## Model Files

| File Name | Quant Type | File Size | File Link | Description |
|-----------|------------|-----------|-----------|-------------|
| ShutterMuse.BF16.gguf | BF16 | 16.4 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.BF16.gguf) | Full BF16 weights. Highest quality, largest file size. |
| ShutterMuse.Q3_K_L.gguf | Q3_K_L | 4.43 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q3_K_L.gguf) | Lower quality but usable, good for low RAM availability. |
| ShutterMuse.Q3_K_M.gguf | Q3_K_M | 4.12 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q3_K_M.gguf) | Low quality. |
| ShutterMuse.Q4_K_M.gguf | Q4_K_M | 5.03 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q4_K_M.gguf) | Good quality, default size for most use cases, *recommended*. |
| ShutterMuse.Q4_K_S.gguf | Q4_K_S | 4.8 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q4_K_S.gguf) | Slightly lower quality with more space savings, *recommended*. |
| ShutterMuse.Q5_K_M.gguf | Q5_K_M | 5.85 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q5_K_M.gguf) | High quality, *recommended*. |
| ShutterMuse.Q5_K_S.gguf | Q5_K_S | 5.72 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q5_K_S.gguf) | High quality, *recommended*. |
| ShutterMuse.Q6_K.gguf | Q6_K | 6.73 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q6_K.gguf) | Very high quality, near perfect, *recommended*. |
| ShutterMuse.Q8_0.gguf | Q8_0 | 8.71 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.Q8_0.gguf) | Extremely high quality, generally unneeded but max available quant. |
| ShutterMuse.mmproj-bf16.gguf | mmproj-bf16 | 1.16 GB | [Link](https://huggingface.co/prithivMLmods/ShutterMuse-GGUF/blob/main/ShutterMuse.mmproj-bf16.gguf) | Multimodal projection file in BF16 format. Used for vision/language models. |

## llama.cpp

LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp