diff --git a/gallery/index.yaml b/gallery/index.yaml index 941f041c67f0..c3e07901dece 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -1,4 +1,29 @@ --- +- name: "qwen-8b-ties-v1" + url: "github:mudler/LocalAI/gallery/virtual.yaml@master" + urls: + - https://huggingface.co/mradermacher/Qwen-8B-Ties-v1-GGUF + description: | + This model, **Qwen-8B-Ties-v1**, is a merged version of the Qwen3-8B base model with additional components from **Alelcv27/Qwen3-8B-Math-CoT** and **Alelcv27/Qwen3-8B-Code**. It uses the **TIES merge method** and is optimized for tasks requiring mathematical reasoning and code execution. The model is designed for large language understanding, with support for both mathematical and programming-related applications. It is a quantized version of the original Qwen3-8B model, but its core architecture remains based on the Qwen3-8B base model. + overrides: + parameters: + model: llama-cpp/models/Qwen-8B-Ties-v1.Q4_K_M.gguf + name: Qwen-8B-Ties-v1-GGUF + backend: llama-cpp + template: + use_tokenizer_template: true + known_usecases: + - chat + function: + grammar: + disable: true + description: Imported from https://huggingface.co/mradermacher/Qwen-8B-Ties-v1-GGUF + options: + - use_jinja:true + files: + - filename: llama-cpp/models/Qwen-8B-Ties-v1.Q4_K_M.gguf + sha256: a02c5752838dd95e03172753b26e12616fc0244148a655e5df30526db2cffc6e + uri: https://huggingface.co/mradermacher/Qwen-8B-Ties-v1-GGUF/resolve/main/Qwen-8B-Ties-v1.Q4_K_M.gguf - name: "glm-4.7-flash-derestricted" url: "github:mudler/LocalAI/gallery/virtual.yaml@master" urls: