40 lines
3.0 KiB
Markdown
40 lines
3.0 KiB
Markdown
---
|
||
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):
|
||
|
||
 |