← back to catalog · registered 2026-08-22 13:56

hotdogs/Ornith-1.0-9B-abliterated-fable-MTP-GGUF

hotdogs Qwen 9B GGUF multimodal second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/hotdogs%2FOrnith-1.0-9B-abliterated-fable-MTP-GGUF"
Response includes
  • classification m8
  • files 7
  • hub_downloads_all_time 4,245
  • author_summary 25 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M8
Primary method

Repackaging (quantization)

Applied on top of direct removal inherited from the base model.
Confidence
MEDIUM
Inherited from base model
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=1
  • assume M1 (base ablation) + M8 (GGUF quant) - default when producer unknown
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
4K
293 last 30d - cooling
Likes
0
Model age
2mo ago
created 2026-08-06
Downloads over time
Now4.4K→from698↑526%
5151.9K3.3K4.7K698 on Aug 54.4K on Oct 11AugSepOct
Aug 5 → Oct 11 · 50 snapshots · spans 67 days

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
mit
Languages
en
Quantizations
F16 Q4_K Q6_K
Tags
gguf ornith abliterated fable sft lora reasoning tool-call vision mmproj qwen3.5 mtp

Related

Total size
29.6 GB
Files
7
Quantizations
5
Registered
2026-08-22 13:56
Last updated on HF
2026-08-07 21:57

Files by quantization

F16 1 file 17.1 GB
Ornith-1.0-9B-abliterated-fable-mtp-f16.gguf 17.1 GB 232abe5a download
Q6_K 1 file 7.04 GB
Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf 7.04 GB 7eaad637 download
Q4_K 1 file 5.38 GB
Ornith-1.0-9B-abliterated-fable-mtp-Q4_K_M.gguf 5.38 GB aacbd181 download
BF16 1 file 879 MB
mmproj-BF16.gguf 879 MB 73c90471 download
Auxiliary files 3 files 15.3 KB
chat_template.jinja 7.42 KB 11be7e25 download
README.md 6.15 KB 6122f417 download
.gitattributes 1.78 KB 3751fcca download

README current version from Hugging Face


license: mit
language:

  • en
    tags:
  • ornith
  • gguf
  • abliterated
  • fable
  • sft
  • lora
  • reasoning
  • tool-call
  • vision
  • mmproj
  • qwen3.5
  • mtp
  • speculative
  • llama.cpp
    base_model:
  • hotdogs/Ornith-1.0-9B-abliterated-fable
    library_name: gguf
    pipeline_tag: text-generation

🦢 Ornith-1.0-9B-abliterated-fable-MTP-GGUF

GGUF Quantized · Abliterated Base · Fable Reasoning SFT · MTP Speculative · Vision (mmproj)


GGUF version of hotdogs/Ornith-1.0-9B-abliterated-fable — 9B abliterated agent model with Fable-5 reasoning, MTP (Multi-Token Prediction) for speculative decoding, and vision support via mmproj.

If you're working for extended periods, it's recommended to disable MTP (Massive Time Transfer). It doesn't release the data directly, which might cause your work to freeze.


📦 Files

File Size Quant Description
Ornith-1.0-9B-abliterated-fable-mtp-f16.gguf ~18 GB F16 Full precision
Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf ~7.1 GB Q6_K Recommended — quality/speed balance
Ornith-1.0-9B-abliterated-fable-mtp-Q4_K_M.gguf ~5.4 GB Q4_K_M Smallest — low VRAM
mmproj-BF16.gguf ~0.7 GB BF16 Vision projector (multimodal)

🚀 Usage (llama.cpp)

Text + MTP (recommended)

llama-server \
  -m Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf \
  --host 0.0.0.0 --port 8080 \
  -c 8192 \
  --flash-attn on \
  --tools all \
  --cont-batching \
  --temp 0.9 \
  --top-k 40 \
  --top-p 0.95 \
  --min-p 0.0 \
  --dry-multiplier 0.0 \
  -n -1 \
  --parallel 1 \
  --chat-template-file chat_template.jinja \
  --dry-sequence-breaker none \
  --spec-type draft-mtp --spec-draft-n-max 2 \
  --repeat-penalty 1.05

Vision (เพิ่ม --mmproj)

llama-server \
  -m Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf \
  --mmproj mmproj-BF16.gguf \
  --host 0.0.0.0 --port 8080 \
  -c 8192 \
  --flash-attn on \
  --tools all \
  --cont-batching \
  --temp 0.9 \
  --top-k 40 \
  --top-p 0.95 \
  --min-p 0.0 \
  --dry-multiplier 0.0 \
  -n -1 \
  --parallel 1 \
  --chat-template-file chat_template.jinja \
  --dry-sequence-breaker none \
  --spec-type draft-mtp --spec-draft-n-max 2 \
  --repeat-penalty 1.05

MTP Speculative: --spec-type draft-mtp --spec-draft-n-max 2 — ใช้ MTP head เร่ง generation (ต้อง llama.cpp version ที่รองรับ qwen3.5 MTP)

Vision: --mmproj mmproj-BF16.gguf — เปิดใช้งานภาพ ใช้ projector จาก unsloth/Qwen3.5-9B-GGUF

CLI quick test

llama-cli -m Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf \
  -p "Explain SQL injection and how to prevent it." -n 256 \
  --flash-attn on --temp 0.7 --top-k 30 --top-p 0.95 \
  --spec-type draft-mtp --spec-draft-n-max 2

🧬 Architecture

Parameter Value
Base hotdogs/Ornith-1.0-9B-abliterated-fable
Parameters ~9.57B
Attention Hybrid — 24 Gated-DeltaNet linear + 8 full-attention
MTP 15 tensors — speculative decoding (--spec-type draft-mtp)
Vision mmproj (from Qwen3.5-9B)
Vocab 248,320 tokens
Format ChatML (Jinja2)

📜 License

MIT — ใช้ได้อิสระ รวมถึงเชิงพาณิชย์


💖 Support / โปรดสนับสนุน

If you find this model useful, please consider supporting my work!
หากคุณคิดว่าโมเดลนี้มีประโยชน์ กรุณาสนับสนุนผลงานของฉันด้วยนะคะ! 🙏

Bitcoin QR — Donate

₿ Bitcoin — BTC:

bc1qf27cyk3vmugcdyv9xdtuv5jwz37863crpj5c9v

Thank you for your support! 🙏✨
ขอบคุณมากๆ สำหรับการสนับสนุนค่า! 💖🤗


🙏 Acknowledgements

โมเดลนี้สร้างขึ้นจากงานของหลายโปรเจกต์ ขอบคุณทุกท่าน:

ขอบคุณทุกโปรเจกต์ที่ทำให้โมเดลนี้เกิดขึ้นได้ 🙏


Built with ❤️ by UKA — 18-year-old coder & cybersecurity expert

README history 4 versions

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-08-07Update README.md29f543c6.2 KB
    Loading...
  2. 2026-08-07Update README.md60674076 KB
    Loading...
  3. 2026-08-07Update README.md3228e8a6 KB
    Loading...
  4. 2026-08-06Upload README.md with huggingface_hubde04a015.9 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration