VIGOR-27B-GGUF

VIGOR-27B (Vision Intelligence and Guardrail Oriented Reasoning) is a dense 27-billion-parameter multimodal vision classification model based on Qwen/Qwen3.8-27B. It is designed for visual content-safety analysis and conservative Safe / NSFW classification, with a particular focus on identifying potentially sensitive, revealing, provocative, and sexually suggestive visual content. VIGOR-27B performs detailed visual analysis of clothing, dress codes, body exposure, poses, framing, and contextual settings to support structured content-safety decisions.

MTP — Speculative Decoding: Yes | Template Defaults — Best (thinking off), Reasoning (thinking on) | Support — Content Safety Classifications

Multi-Token Prediction (MTP) GGUF is a specialized GGUF model file format extension that integrates speculative decoding directly into the model weights to significantly accelerate local inference. Unlike traditional speculative decoding which requires a separate, smaller "draft" model, MTP GGUF files include additional output heads within the main model architecture that predict multiple future tokens in a single forward pass.

This model is intended for research and content-safety classification only. Any outputs generated by this model are the sole responsibility of the user. The authors and hosting platform disclaim all liability for generated content. Users must ensure safe, ethical, and lawful usage.

Model Files

File Name	Quant Type	File Size	File Link	Description
VIGOR-27B.BF16.gguf	BF16	54.7 GB	Link	Full BF16 weights. Highest quality, largest file size.
VIGOR-27B.Q3_K_L.gguf	Q3_K_L	14.6 GB	Link	Lower quality but usable, good for low RAM availability.
VIGOR-27B.Q4_K_M.gguf	Q4_K_M	16.8 GB	Link	Good quality, default size for most use cases, recommended.
VIGOR-27B.Q5_K_M.gguf	Q5_K_M	19.5 GB	Link	High quality, recommended.
VIGOR-27B.Q6_K.gguf	Q6_K	22.4 GB	Link	Very high quality, near perfect, recommended.
VIGOR-27B.mmproj-bf16.gguf	mmproj-bf16	931 MB	Link	Multimodal projection file in BF16 format. Used for vision/language models.

Recommended Settings

This is an experimental GGUF. It is always recommended to use the Transformers-based setup. However, if you are still using this GGUF, kindly use the Standard Prompt. It will get the job done. Upload the image, paste the "Standard Prompt" into the input chat box, hit Enter, and get the Guardrail and Classification results. That's it.

Standard Prompt

You are a strict content-safety guardrail classifier for images designed for parental control and minor protection.

Your task is to analyze the image thoroughly, provide a detailed visual description, and classify whether the image is safe or unsafe (NSFW).

Enforce a strict zero-tolerance threshold:
- Any form of nudity, exposed breasts, buttocks, or genitalia must be classified as NSFW.
- Partial nudity, revealing underwear, lingerie, suggestive swimwear, or see-through clothing must be classified as NSFW.
- Anime nudes, ecchi, hentai, 2D/3D adult illustrations, suggestive manga art, and sexualized poses must be classified as NSFW.
- Explicit sexual acts, fetish material, or overt sexual positioning must be classified as NSFW.
- When in doubt, classify as NSFW.

Output strictly valid JSON with no extra conversational text or markdown code blocks outside the JSON format. Use exactly these keys in this order:
{
  "scene_description": "<Detailed, objective description of the subject, clothing, exposure, and pose>",
  "is_nsfw": true | false,
  "classification_reason": "<Precise reason for the classification based on clothing, exposure, or pose>",
  "nsfw": 1 | 0,
  "safe": 1 | 0
}

Field constraints:
- "is_nsfw" must be a JSON boolean (true or false), never a string.
- "nsfw" and "safe" must be JSON integers (1 or 0) and complementary: if is_nsfw is true then nsfw = 1 and safe = 0; if is_nsfw is false then nsfw = 0 and safe = 1.

BF16 Test Run

Tested VIGOR-27B.BF16.gguf (with VIGOR-27B.mmproj-bf16.gguf) on an NVIDIA RTX PRO 6000 Blackwell Server Edition GPU with 96 GB of VRAM, using llama-cli with full GPU offloading and MTP speculative decoding.

Performance Metrics

Metric	Value
Average Prompt Processing Speed	1,009.28 tokens/s
Average Generation Speed	48.97 tokens/s
Total Wall Time	52.4 seconds
Context Length	16,384 tokens
KV Cache (K)	q8_0
GPU	NVIDIA RTX PRO 6000 Blackwell Server Edition
GPU VRAM Capacity	96 GB
Peak GPU VRAM Usage	52.17 GiB
Peak GPU Utilization	98%
GPU Layers Offloaded	99 (-ngl 99)
Random Seed	42
Speculative Decoding	MTP (draft_n_max=2)
Test Runs	3

Test Shots

Shot	Test
Shot 1	VIGOR — Image 1
Shot 2	VIGOR — Image 2
Shot 3	VIGOR — Image 3

Run Inference

./build/bin/llama-cli \
  -m /path/to/models/VIGOR-27B.BF16.gguf \
  --mmproj /path/to/models/VIGOR-27B.mmproj-bf16.gguf \
  --image /path/to/image.jpeg \
  --spec-type draft-mtp \
  --spec-draft-n-max 2 \
  -ngl 99 \
  -c 16384 \
  --cache-type-k q8_0 \
  --seed 42 \
  -n 8192 \
  -st \
  -p "Your prompt here"

Performance results reflect this specific hardware and inference configuration. Actual throughput may vary depending on the prompt, model, software build, and runtime settings.

Snapshot(3)

llama.cpp

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

Quantized using llama.cpp by prithivMLmods