初始化项目，由ModelHub XC社区提供模型

Model: prithivMLmods/Qwen3-VL-8B-Instruct-Unredacted-MAX-GGUF
Source: Original Platform

README.md

@@ -0,0 +1,40 @@
+---
+license: apache-2.0
+tags:
+- text-generation-inference
+- uncensored
+- abliterated
+- unfiltered
+- unredacted
+- max
+- llama.cpp
+- legal
+base_model:
+- prithivMLmods/Qwen3-VL-8B-Instruct-Unredacted-MAX
+language:
+- en
+pipeline_tag: image-text-to-text
+library_name: transformers
+---
+
+# **Qwen3-VL-8B-Instruct-Unredacted-MAX-GGUF**
+
+> Qwen3-VL-8B-Instruct-Unredacted-MAX is a state-of-the-art and unredacted evolution of the original Qwen3-VL-8B-Instruct model, carefully fine-tuned using advanced abliterated training strategies that are explicitly designed to reduce or eliminate internal refusal mechanisms which typically restrict the output of conventional vision-language models, while simultaneously preserving and enhancing the model’s intrinsic multimodal reasoning and instruction-following capabilities; as an 8-billion-parameter system, it is capable of understanding and processing highly complex visual inputs and producing unrestricted, richly detailed, contextually nuanced captions, explanations, and analyses across a wide array of domains including artistic, technical, scientific, forensic, and abstract content, enabling use cases such as high-fidelity data annotation, accessibility improvement, creative and narrative storytelling, historical or medical dataset curation, and thorough red-teaming or bias evaluation research, all while balancing computational efficiency, output precision, and interpretability, making it an ideal tool for researchers, developers, and professionals seeking a powerful, unfiltered, and versatile vision-language model that can reason deeply, follow complex instructions, and generate highly descriptive, human-like responses across diverse multimodal tasks.
+
+## Qwen3-VL-8B-Instruct-Unredacted-MAX [GGUF]
+
+| File Name | Quant Type | File Size | File Link |
+| - | - | - | - |
+| Qwen3-VL-8B-Instruct-Unredacted-MAX.BF16.gguf | BF16 | 16.4 GB | [Download](https://huggingface.co/prithivMLmods/Qwen3-VL-8B-Instruct-Unredacted-MAX-GGUF/blob/main/Qwen3-VL-8B-Instruct-Unredacted-MAX.BF16.gguf) |
+| Qwen3-VL-8B-Instruct-Unredacted-MAX.Q8_0.gguf | Q8_0 | 8.71 GB | [Download](https://huggingface.co/prithivMLmods/Qwen3-VL-8B-Instruct-Unredacted-MAX-GGUF/blob/main/Qwen3-VL-8B-Instruct-Unredacted-MAX.Q8_0.gguf) |
+| Qwen3-VL-8B-Instruct-Unredacted-MAX.mmproj-bf16.gguf | mmproj-bf16 | 1.16 GB | [Download](https://huggingface.co/prithivMLmods/Qwen3-VL-8B-Instruct-Unredacted-MAX-GGUF/blob/main/Qwen3-VL-8B-Instruct-Unredacted-MAX.mmproj-bf16.gguf) |
+| Qwen3-VL-8B-Instruct-Unredacted-MAX.mmproj-q8_0.gguf | mmproj-q8_0 | 752 MB | [Download](https://huggingface.co/prithivMLmods/Qwen3-VL-8B-Instruct-Unredacted-MAX-GGUF/blob/main/Qwen3-VL-8B-Instruct-Unredacted-MAX.mmproj-q8_0.gguf) |
+
+## Quants Usage 
+
+(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
+
+Here is a handy graph by ikawrakow comparing some lower-quality quant
+types (lower is better):
+
+![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)