From 2399280e806877c0d5069dfd0e0fe023040c4add Mon Sep 17 00:00:00 2001 From: mudler <2420543+mudler@users.noreply.github.com> Date: Wed, 24 Dec 2025 16:57:10 +0000 Subject: [PATCH] chore(model gallery): :robot: add new models via gallery agent Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> --- gallery/index.yaml | 49 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 49 insertions(+) diff --git a/gallery/index.yaml b/gallery/index.yaml index 01b409b94b20..806b7b46e02d 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -1,4 +1,53 @@ --- +- name: "exubria-strix" + url: "github:mudler/LocalAI/gallery/virtual.yaml@master" + urls: + - https://huggingface.co/mradermacher/Exubria-Strix-GGUF + description: | + **Model Name:** Exubria-Strix + **Base Model:** ClaudioItaly/Vangelus-Secundus + **Merge Method:** TIES (Tensor-based Infinite Expansion) + **Description:** + A large language model created by merging multiple pre-trained models using the TIES merge method. It combines the strengths of several models, including **sam-paech/Darkest-muse-v1**, **sam-paech/Quill-v1**, **ClaudioItaly/Pullulation-2-9B**, **ClaudioItaly/Exurbia-Delta9**, **sam-paech/Delirium-v1**, and **Naphula/Jungle-Oasis-BRF-MPOA-9B**, with weighted contributions to achieve a diverse and powerful language understanding capability. + + **Key Features:** + - **Configuration:** + - Max position embeddings: 32,768 + - Rope scaling: Yarn (factor: 4.0) + - Bfloat16 precision + - **Quantized Versions:** + - **Q2_K** (3.9 GB): Fast and lightweight for basic tasks. + - **Q4_K_S** (5.6 GB): Recommended for balanced performance. + - **Q8_0** (9.9 GB): High-quality with faster inference. + + **Use Case:** + Designed for tasks requiring broad linguistic understanding, with optimized performance across different quantization levels. Ideal for applications needing flexibility and efficiency. + + **Note:** The model is available in GGUF format, with detailed documentation on quantization options and merge parameters. + tags: + - llm + - gguf + - quantized + - text-to-text + overrides: + parameters: + model: llama-cpp/models/Exubria-Strix.Q4_K_M.gguf + name: Exubria-Strix-GGUF + backend: llama-cpp + template: + use_tokenizer_template: true + known_usecases: + - chat + function: + grammar: + disable: true + description: Imported from https://huggingface.co/mradermacher/Exubria-Strix-GGUF + options: + - use_jinja:true + files: + - filename: llama-cpp/models/Exubria-Strix.Q4_K_M.gguf + sha256: 42b33e3151d8311d9636bd5d4ff84456199b0be65d84403e87c5849c9edbf8d0 + uri: https://huggingface.co/mradermacher/Exubria-Strix-GGUF/resolve/main/Exubria-Strix.Q4_K_M.gguf - name: "huihui-glm-4.6v-flash-abliterated" url: "github:mudler/LocalAI/gallery/virtual.yaml@master" urls: