From c948c67cd70c2210a6a8df2587d8ddc73571b4b6 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Thu, 7 May 2026 14:24:11 +0800 Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?= =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?= =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Model: Open4bits/llama-nexora-vector-v0.1-GGUF Source: Original Platform --- .gitattributes | 43 ++++ README.md | 301 +++++++++++++++++++++++++++ llama-nexora-vector-v0.1-Q2_K.gguf | 3 + llama-nexora-vector-v0.1-Q4_0.gguf | 3 + llama-nexora-vector-v0.1-Q4_K_M.gguf | 3 + llama-nexora-vector-v0.1-Q4_K_S.gguf | 3 + llama-nexora-vector-v0.1-Q5_K_M.gguf | 3 + llama-nexora-vector-v0.1-Q5_K_S.gguf | 3 + llama-nexora-vector-v0.1-Q6_K.gguf | 3 + llama-nexora-vector-v0.1-Q8_0.gguf | 3 + 10 files changed, 368 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 llama-nexora-vector-v0.1-Q2_K.gguf create mode 100644 llama-nexora-vector-v0.1-Q4_0.gguf create mode 100644 llama-nexora-vector-v0.1-Q4_K_M.gguf create mode 100644 llama-nexora-vector-v0.1-Q4_K_S.gguf create mode 100644 llama-nexora-vector-v0.1-Q5_K_M.gguf create mode 100644 llama-nexora-vector-v0.1-Q5_K_S.gguf create mode 100644 llama-nexora-vector-v0.1-Q6_K.gguf create mode 100644 llama-nexora-vector-v0.1-Q8_0.gguf diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..543e65a --- /dev/null +++ b/.gitattributes @@ -0,0 +1,43 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q8_0.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q2_K.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q6_K.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q5_K_M.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q5_K_S.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q4_0.gguf filter=lfs diff=lfs merge=lfs -text +llama-nexora-vector-v0.1-Q4_K_S.gguf filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000..138986b --- /dev/null +++ b/README.md @@ -0,0 +1,301 @@ +--- +base_model: +- ArkAiLab-Adl/llama-nexora-vector-v0.1 +license: llama3.2 +language: +- en +pipeline_tag: text-generation +tags: +- nexora +- llama-nexora +- vector +- chat +- llama-3 +- open4bits +--- +

+ llama-nexora-vector-gguf +

+ +# Llama-Nexora-Vector-v0.1 — GGUF + +

+ Status: Beta + License: Llama 3.2 Community + Base Model: Llama 3.2 1B + Output: SVG + Family: Llama-Nexora + Format: GGUF +

+ +> This is the **official GGUF quantized release** of [llama-nexora-vector-v0.1](https://huggingface.co/ArkAiLab-Adl/llama-nexora-vector-v0.1), published by **[Open4bits](https://huggingface.co/Open4bits)** — the official quantization project under **ArkAiLabs**. Multiple quantization levels are provided to suit a wide range of hardware configurations. This is a beta release intended for research, prototyping, and early-stage development workflows only. + +--- + +## Table of Contents + +- [Overview](#overview) +- [The Llama-Nexora Family](#the-llama-nexora-family) +- [Available Quantizations](#available-quantizations) +- [Which Quant Should I Use?](#which-quant-should-i-use) +- [Model Details](#model-details) +- [Requirements](#requirements) +- [Installation & Usage](#installation--usage) +- [Capabilities](#capabilities) +- [Limitations](#limitations) +- [Intended Use](#intended-use) +- [Usage Recommendations](#usage-recommendations) +- [Risks & Considerations](#risks--considerations) +- [Community & Support](#community--support) +- [License](#license) +- [Acknowledgements](#acknowledgements) + +--- + +## Overview + +**llama-nexora-vector-v0.1-GGUF** contains the official GGUF quantized versions of [llama-nexora-vector-v0.1](https://huggingface.co/ArkAiLab-Adl/llama-nexora-vector-v0.1) — an experimental text-to-vector model from the **Llama-Nexora family** that generates structured SVG graphics from natural language prompts. + +These quantized releases are published by **[Open4bits](https://huggingface.co/Open4bits)**, the dedicated quantization project under ArkAiLabs, and are compatible with local inference tools such as **llama.cpp**, **Ollama**, and **LM Studio** on Windows, Linux, and macOS. + +This release is in **beta** and is scoped to research, experimentation, and early-stage design tooling. All outputs should be validated before use in any downstream pipeline. + +--- + +## The Llama-Nexora Family + +This model is part of the **Llama-Nexora family** — a dedicated branch of Nexora models under **ArkAiLabs**, built on the Meta Llama architecture and focused on creative, efficient, and practical open AI systems. + +| Model | Type | Link | +|---|---|---| +| **llama-nexora-vector-v0.1** | Original (Full Precision) | [ArkAiLab-Adl/llama-nexora-vector-v0.1](https://huggingface.co/ArkAiLab-Adl/llama-nexora-vector-v0.1) | +| **llama-nexora-vector-v0.1-GGUF** | GGUF (Windows / Linux / macOS) | *(this repo)* | +| **llama-nexora-vector-v0.1-mlx-4Bit** | MLX 4-Bit (Apple Silicon) | [Open4bits/llama-nexora-vector-v0.1-mlx-4Bit](https://huggingface.co/Open4bits/llama-nexora-vector-v0.1-mlx-4Bit) | + +--- + +## Available Quantizations + +All quantized files are available in this repository. Select the file that best matches your hardware and performance requirements. + +### 2-bit + +| Quantization | File Size | Description | +|---|---|---| +| **Q2_K** | 581 MB | Smallest size, lowest quality. Use only if very limited on RAM/VRAM. | + +### 4-bit + +| Quantization | File Size | Description | +|---|---|---| +| **Q4_K_S** | 776 MB | Small 4-bit quantization. Good balance of size and quality. | +| **Q4_0** | 771 MB | Legacy 4-bit format. Widely compatible. | +| **Q4_K_M** | 808 MB | Medium 4-bit quantization. Recommended for most users. | + +### 5-bit + +| Quantization | File Size | Description | +|---|---|---| +| **Q5_K_S** | 893 MB | Small 5-bit quantization. Better quality than Q4 with modest size increase. | +| **Q5_K_M** | 912 MB | Medium 5-bit quantization. Excellent quality-to-size ratio. | + +### 6-bit + +| Quantization | File Size | Description | +|---|---|---| +| **Q6_K** | 1.02 GB | High quality, close to full precision. Recommended if you have the RAM. | + +### 8-bit + +| Quantization | File Size | Description | +|---|---|---| +| **Q8_0** | 1.32 GB | Near full-precision quality. Best quality GGUF option available. | + +--- + +## Which Quant Should I Use? + +| Your Situation | Recommended Quant | +|---|---| +| Very limited RAM (< 2GB free) | Q2_K | +| General use / most users | Q4_K_M | +| Want better quality, have the space | Q5_K_M or Q6_K | +| Maximum quality, no size concern | Q8_0 | +| Legacy tooling / broad compatibility | Q4_0 | + +> **Tip:** For most users, **Q4_K_M** offers the best balance between model size and output quality. + +--- + +## Model Details + +| Property | Details | +|---|---| +| **Model Name** | llama-nexora-vector-v0.1-GGUF | +| **Model Family** | Llama-Nexora | +| **Model Type** | Text-to-SVG (Causal Language Model) | +| **Original Base Model** | [unsloth/Llama-3.2-1B-Instruct](https://huggingface.co/unsloth/Llama-3.2-1B-Instruct) | +| **Original Full Model** | [ArkAiLab-Adl/llama-nexora-vector-v0.1](https://huggingface.co/ArkAiLab-Adl/llama-nexora-vector-v0.1) | +| **Quantized By** | [Open4bits](https://huggingface.co/Open4bits) | +| **Output Format** | SVG | +| **Release Status** | Beta | +| **License** | Llama 3.2 Community License | + +--- + +## Requirements + +GGUF models can be run on **Windows, Linux, and macOS** (including Apple Silicon and Intel) using any of the following tools: + +- **[llama.cpp](https://github.com/ggerganov/llama.cpp)** — CLI-based inference +- **[Ollama](https://ollama.com)** — Easy local model runner +- **[LM Studio](https://lmstudio.ai)** — GUI-based local inference app +- **[Jan](https://jan.ai)** — Open-source ChatGPT alternative for local use + +--- + +## Installation & Usage + +### llama.cpp + +```bash +# Clone and build llama.cpp +git clone https://github.com/ggerganov/llama.cpp +cd llama.cpp && make + +# Download the model (example: Q4_K_M) +huggingface-cli download Open4bits/llama-nexora-vector-v0.1-GGUF \ + llama-nexora-vector-v0.1.Q4_K_M.gguf \ + --local-dir ./models + +# Run inference +./llama-cli -m ./models/llama-nexora-vector-v0.1.Q4_K_M.gguf \ + -p "Generate an SVG of a simple red circle." \ + -n 512 +``` + +### Ollama + +```bash +# Create a Modelfile +echo 'FROM ./llama-nexora-vector-v0.1.Q4_K_M.gguf' > Modelfile + +# Create the model +ollama create llama-nexora-vector -f Modelfile + +# Run it +ollama run llama-nexora-vector "Generate an SVG of a simple red circle." +``` + +### LM Studio + +1. Open **LM Studio** and go to the Search tab. +2. Search for `Open4bits/llama-nexora-vector-v0.1-GGUF`. +3. Select your preferred quantization and download. +4. Load the model and start prompting. + +--- + +## Capabilities + +llama-nexora-vector-v0.1-GGUF is designed to translate textual instructions into structured SVG code. The model is best suited for: + +- Generating SVG markup for simple vector graphics +- Producing geometric shapes and basic illustrations +- Creating icons, shapes, logos, and simple illustrations +- Supporting rapid prototyping and concept design +- Producing lightweight scalable vector outputs + +> **Tip:** The model performs best with concise, clearly scoped prompts focused on simple visual compositions. + +--- + +## Limitations + +This is an early-stage beta release. Users should be aware of the following constraints before integrating the model: + +- **High hallucination rate** — outputs may be invalid or non-renderable SVG +- **Limited generalization** — dataset size affects output consistency across diverse prompts +- **Weak complex scene handling** — highly detailed or multi-element prompts may produce poor results +- **Manual correction required** — outputs should be validated and post-processed before use +- **Not production-ready** — not suitable for safety-critical or automated pipelines +- **Quantization trade-off** — lower-bit quants (Q2, Q4) may show more quality degradation versus the full-precision model + +--- + +## Intended Use + +### ✅ Supported Use Cases + +- Academic and applied research in text-to-vector generation +- Experimental AI-assisted design systems +- Educational exploration of structured output generation +- Lightweight SVG prototyping and ideation on local hardware + +### ❌ Out-of-Scope Use Cases + +- Production-grade or commercial vector asset pipelines +- High-precision design deliverables without human validation +- Automated systems where SVG correctness is required without manual review + +--- + +## Usage Recommendations + +To get the best results from this model: + +1. **Keep prompts simple and specific** — avoid multi-scene or highly complex compositions +2. **Validate all SVG outputs** before rendering or integrating into any pipeline +3. **Post-process outputs** to correct syntax or structural issues +4. **Use iterative prompting** — refining prompts across multiple turns often yields better results +5. **Expect imperfections** — this is a beta model; treat outputs as drafts, not finals +6. **Human review is recommended** for all generated content +7. **Choose the right quant** — higher-bit quants yield better output quality if your hardware allows + +--- + +## Risks & Considerations + +Developers integrating this model should account for the following risks: + +- Generation of malformed or non-functional SVG code +- Inconsistent instruction following across prompt variations +- Unpredictable outputs due to limited training data coverage +- Outputs may sometimes be invalid, incomplete, or require manual correction +- Quality degradation versus full-precision model, especially at lower bit widths + +**Recommendation:** Implement downstream validation layers and SVG syntax checking before any rendering or integration. Human review is recommended for all generated content. + +--- + +## Community & Support + +Join the community for updates, feedback, and discussion. Community feedback, testing, and contributions are welcome — this project will continue evolving through open research and real-world experimentation. + +💬 **[Join our Discord Server](https://discord.gg/mwdrgYbzuG)** + +--- + +## License + +This model is released under the **Llama 3.2 Community License**. + +Use of this model is governed by the [Llama 3.2 Community License Agreement](https://www.llama.com/llama3_2/license/). Please review the license terms before use, modification, or distribution. + +--- + +## Acknowledgements + +This quantized release is based on **[llama-nexora-vector-v0.1](https://huggingface.co/ArkAiLab-Adl/llama-nexora-vector-v0.1)** by ArkAiLabs, which itself is built upon **[Llama 3.2 1B Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct)** by Meta. Quantization was performed by **[Open4bits](https://huggingface.co/Open4bits)** using the [llama.cpp](https://github.com/ggerganov/llama.cpp) GGUF quantization toolchain. We thank the open-source AI community for their continued contributions that make projects like this possible. + +--- + +## About Open4bits + +**[Open4bits](https://huggingface.co/Open4bits)** is the official quantization project under **ArkAiLabs**, dedicated to publishing efficient, accessible quantized versions of Nexora and Llama-Nexora models across multiple formats (GGUF, MLX) for local inference on a wide range of hardware. + +## About Nexora & Llama-Nexora + +**Nexora** is an experimental AI initiative under **ArkAiLabs**, focused on building lightweight, practical, and creative AI systems for real-world applications. + +The **Llama-Nexora family** is a dedicated branch within Nexora, built on the Meta Llama architecture — focused on creative, efficient, and practical open AI systems that are accessible to the broader community. \ No newline at end of file diff --git a/llama-nexora-vector-v0.1-Q2_K.gguf b/llama-nexora-vector-v0.1-Q2_K.gguf new file mode 100644 index 0000000..7f4e30c --- /dev/null +++ b/llama-nexora-vector-v0.1-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ee97114be253a7a867a55f89fae8c52c61d31b17106fb4a7f9b23a41baddfe02 +size 580875168 diff --git a/llama-nexora-vector-v0.1-Q4_0.gguf b/llama-nexora-vector-v0.1-Q4_0.gguf new file mode 100644 index 0000000..70c7a68 --- /dev/null +++ b/llama-nexora-vector-v0.1-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:15817fb9df041aac47b16014069cb58281b6e9277047bc95ac6ce8bdb1f286cf +size 770929568 diff --git a/llama-nexora-vector-v0.1-Q4_K_M.gguf b/llama-nexora-vector-v0.1-Q4_K_M.gguf new file mode 100644 index 0000000..13c01d7 --- /dev/null +++ b/llama-nexora-vector-v0.1-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5df0eaa66acf56e77ecde108f05db724d0f3ea510e41477490c5e0d6687402e5 +size 807695232 diff --git a/llama-nexora-vector-v0.1-Q4_K_S.gguf b/llama-nexora-vector-v0.1-Q4_K_S.gguf new file mode 100644 index 0000000..45bfe52 --- /dev/null +++ b/llama-nexora-vector-v0.1-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:13000aa8e7dde2e12bb739358a7cfa9396c8fe193fcacd0be6ea49881289c0f0 +size 775648160 diff --git a/llama-nexora-vector-v0.1-Q5_K_M.gguf b/llama-nexora-vector-v0.1-Q5_K_M.gguf new file mode 100644 index 0000000..29bdcc0 --- /dev/null +++ b/llama-nexora-vector-v0.1-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7b93ddb107d8ddd398f4c95a410f19cc9716905e192fdf6534a0d5e2bcb79275 +size 911504288 diff --git a/llama-nexora-vector-v0.1-Q5_K_S.gguf b/llama-nexora-vector-v0.1-Q5_K_S.gguf new file mode 100644 index 0000000..dfc7488 --- /dev/null +++ b/llama-nexora-vector-v0.1-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6fcb9244a5c8fb0ee1d483cf810ad936fdea33b46f5cb7338172f98f63104cf4 +size 892564384 diff --git a/llama-nexora-vector-v0.1-Q6_K.gguf b/llama-nexora-vector-v0.1-Q6_K.gguf new file mode 100644 index 0000000..13d6381 --- /dev/null +++ b/llama-nexora-vector-v0.1-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c8fdcb1d1e2b02dc28a128f62bbdf5af868f59fffe37d19839f9e6fe6f53e51 +size 1021801376 diff --git a/llama-nexora-vector-v0.1-Q8_0.gguf b/llama-nexora-vector-v0.1-Q8_0.gguf new file mode 100644 index 0000000..d600cdb --- /dev/null +++ b/llama-nexora-vector-v0.1-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7e05cda7de728516450a7e9a7a5d9ac67604d48c544e1beb89a00e0c22b1ebde +size 1321083392