Views
No views yet
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.
| 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) |
1llama-server \
2 -m Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf \
3 --host 0.0.0.0 --port 8080 \
4 -c 8192 \
5 --flash-attn on \
6 --tools all \
7 --cont-batching \
8 --temp 0.9 \
9 --top-k 40 \
10 --top-p 0.95 \
11 --min-p 0.0 \
12 --dry-multiplier 0.0 \
13 -n -1 \
14 --parallel 1 \
15 --chat-template-file chat_template.jinja \
16 --dry-sequence-breaker none \
17 --spec-type draft-mtp --spec-draft-n-max 2 \
18 --repeat-penalty 1.051llama-server \
2 -m Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf \
3 --mmproj mmproj-BF16.gguf \
4 --host 0.0.0.0 --port 8080 \
5 -c 8192 \
6 --flash-attn on \
7 --tools all \
8 --cont-batching \
9 --temp 0.9 \
10 --top-k 40 \
11 --top-p 0.95 \
12 --min-p 0.0 \
13 --dry-multiplier 0.0 \
14 -n -1 \
15 --parallel 1 \
16 --chat-template-file chat_template.jinja \
17 --dry-sequence-breaker none \
18 --spec-type draft-mtp --spec-draft-n-max 2 \
19 --repeat-penalty 1.05MTP 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
1llama-cli -m Ornith-1.0-9B-abliterated-fable-mtp-Q6_K.gguf \
2 -p "Explain SQL injection and how to prevent it." -n 256 \
3 --flash-attn on --temp 0.7 --top-k 30 --top-p 0.95 \
4 --spec-type draft-mtp --spec-draft-n-max 2| 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) |
bc1qf27cyk3vmugcdyv9xdtuv5jwz37863crpj5c9v