---

base_model: simpledirect/Vinci-Piccolo-1.0

library_name: gguf

pipeline_tag: text-generation

tags:

- gguf

- quantized

- qwen3_5

- conversational

- chat

- character

- simpledirect

- vinci

license: apache-2.0

widget:

• messages:

- role: user

content: "What's a hard truth I need to hear?"

• messages:

- role: user

content: "Who are you?"

---

<img src="./assets/vinci-vero-header.svg" alt="Vinci with Vero, its character" width="280" height="64">

Vinci Piccolo 1.0 — GGUF

GGUF (quantized) builds of Vinci Piccolo, for local inference with Ollama, LM Studio, and llama.cpp. For the full-precision weights, evals, and details, see simpledirect/Vinci-Piccolo-1.0.

These files are not the measured artifact

🔴 No GGUF file in this repository has been evaluated. The benchmark figures on the full-precision card were measured on the unquantized bf16 weights, not on any file here. Quantization shifts scores, and with no per-tier measurement this card cannot say in which direction or by how much for any tier.

This card therefore publishes no per-tier score, no quality ranking between the tiers, and no recommended tier. The file sizes and memory figures below are measured, and describe storage and memory cost only. A file's presence in this repository is not a performance claim about that tier, and not a claim of parity with the full-precision weights.

Available variants

| File | Size | Min RAM | Notes |

|---|---|---|---|

| vinci-piccolo-1.0-20260629-Q6_K.gguf | 3.46 GB | 12 GB | 6-bit K-quant |

| vinci-piccolo-1.0-20260629-Q5_K_M.gguf | 3.07 GB | 10 GB | 5-bit K-quant, medium mixture |

| vinci-piccolo-1.0-20260629-Q4_K_M.gguf | 2.71 GB | 8 GB | 4-bit K-quant, medium mixture |

GPU: Q5_K_M and Q4_K_M run on 4 GB VRAM; Q6_K needs 6 GB. Mac M-series: Q5_K_M fits on 8 GB unified memory; Q6_K needs 16 GB.

Ollama

ollama run hf.co/simpledirect/Vinci-Piccolo-1.0-GGUF

llama.cpp

./llama-cli \
    -m vinci-piccolo-1.0-20260629-Q5_K_M.gguf \
    --ctx-size 262144 \
    --temp 0 \
    --chat-template qwen3

llama-server (OpenAI-compatible API)

./llama-server \
    -m vinci-piccolo-1.0-20260629-Q5_K_M.gguf \
    --ctx-size 262144 \
    --host 0.0.0.0 \
    --port 8080

Prompt format

Qwen / ChatML chat template. No system prompt required — character is trained into the weights. Pass enable_thinking=False when using the tokenizer directly to suppress <think> output.

Citation

@misc{simpledirect2026vinci,
  title        = {Vinci Piccolo 1.0},
  author       = {{SimpleDirect}},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/simpledirect/Vinci-Piccolo-1.0}},
  note         = {Apache 2.0. Fine-tuned from Qwen/Qwen3.5-4B.},
}