From da02b17be3ee3c6cc7f82bb47d773709c479f351 Mon Sep 17 00:00:00 2001 From: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com> Date: Sat, 12 Sep 2026 12:05:08 +0000 Subject: [PATCH] feat(gallery): add Spark 4B Q8 and remove duplicate Offer the official Spark-X2.5-4B Q8 build as a variant of the existing Q4 entry. Pin both downloads and verify their SHA256 values against Hugging Face LFS metadata and download headers. Document installation. Remove the duplicate Qwopus entry that points at a DeepSeek repository page without a checksum. Keep the valid Qwopus Q4 and Q8 entries. Assisted-by: Codex:gpt-6 --- docs/content/features/model-gallery.md | 10 ++- gallery/index.yaml | 101 +++++++++++-------------- 2 files changed, 52 insertions(+), 59 deletions(-) diff --git a/docs/content/features/model-gallery.md b/docs/content/features/model-gallery.md index 4cd6f719a..d19ef8a9c 100644 --- a/docs/content/features/model-gallery.md +++ b/docs/content/features/model-gallery.md @@ -43,7 +43,7 @@ Both views use the same model selection and store the view, search, filter, and selection in the URL. Installing from Explore does not move you away from the catalog; the entry updates in place when the operation finishes. -## Spark-X2.5-1.7B +## Spark-X2.5 Install Spark-X2.5-1.7B with automatic selection between its Q4_K_M and Q8_0 GGUF builds: @@ -58,6 +58,14 @@ To select the Q8_0 build explicitly: local-ai models install spark-x2.5-1.7b-q4 --variant spark-x2.5-1.7b-q8 ``` +Spark-X2.5-4B also offers Q4_K_M and Q8_0 builds. Install with automatic +variant selection, or select Q8_0 explicitly: + +```bash +local-ai models install spark-x2.5-4b +local-ai models install spark-x2.5-4b --variant spark-x2.5-4b-q8 +``` + These text-only builds use the llama.cpp backend and the embedded Jinja chat template. The gallery defaults to a 32,768-token context to limit memory use. The [source model](https://huggingface.co/XHToken/Spark-X2.5-1.7B) supports up to diff --git a/gallery/index.yaml b/gallery/index.yaml index 5b0139e72..7e283f77d 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -1,35 +1,29 @@ --- -- name: "spark-x2.5-4b" +- &spark-x2-5-4b + name: "spark-x2.5-4b" + variants: + - model: spark-x2.5-4b-q8 url: "github:mudler/LocalAI/gallery/virtual.yaml@master" urls: + - https://huggingface.co/XHToken/Spark-X2.5-4B - https://huggingface.co/XHToken/Spark-X2.5-4B-GGUF description: | - # Spark-X2.5 - - [](https://join.slack.com/t/tokenspark/shared_invite/zt-432qf8l2f-5~dLyXv8uETr0P0UuC07nw) - [](https://discord.gg/kTDE2Hg8aw) - [](https://www.youtube.com/@SparkLLM) - [](https://dev.to/sparkllm) - [](https://bsky.app/profile/sparkllm.bsky.social) - [](https://x.com/sparkllm) - [](https://www.zhihu.com/people/zhiikz7qh7m) - [](images/xhtoken-wechat.jpg) - - > [!Note] - > This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format. - - ## Introduction - - We are introducing Spark-X2.5-4B and Spark-X2.5-1.7B, two compact, general-purpose language models designed to make capable AI more practical, efficient, and accessible. The models deliver strong performance across a broad range of everyday tasks—including conversation, writing, translation, reasoning, coding, tool use, and agentic workflows—achieving leading results among open-source models of comparable size. Spark-X2.5 combines an efficiency-oriented architecture with native context windows of up to 1M tokens, and support for more than 200 languages. - - ... + Spark-X2.5-4B is XHToken's 4B text model for conversation, reasoning, + coding, and multilingual tasks. This build uses Q4_K_M GGUF weights, + the embedded Jinja chat template, and a 32K-token default context. license: "apache-2.0" tags: - llm - gguf - icon: https://img.shields.io/badge/WeChat-Join-07C160?logo=wechat&logoColor=white)](images/xhtoken-wechat.jpg + - cpu + - gpu + - reasoning + - coding + - multilingual + last_checked: "2026-09-12" overrides: backend: llama-cpp + context_size: 32768 function: automatic_tool_parsing_fallback: true grammar: @@ -40,59 +34,50 @@ - use_jinja:true parameters: model: llama-cpp/models/Spark-X2.5-4B-Q4_K_M/Spark-X2.5-4B-Q4_K_M.gguf + temperature: 1 + top_p: 0.95 + top_k: -1 + min_p: 0 + repeat_penalty: 1 template: use_tokenizer_template: true files: - filename: llama-cpp/models/Spark-X2.5-4B-Q4_K_M/Spark-X2.5-4B-Q4_K_M.gguf sha256: adfcfa19a4ed6a5985da8bf565fe15f8e1a7e131d79bae2d19d48d1c40109428 - uri: https://huggingface.co/XHToken/Spark-X2.5-4B-GGUF/resolve/main/Spark-X2.5-4B-Q4_K_M.gguf -- name: "qwopus3.8-27b-flash" - url: "github:mudler/LocalAI/gallery/virtual.yaml@master" - urls: - - https://huggingface.co/unsloth/DeepSeek-V4-Flash-Vision-Exp-GGUF + uri: https://huggingface.co/XHToken/Spark-X2.5-4B-GGUF/resolve/d313463f1ff7dc29db20193fddc4ad8386d16261/Spark-X2.5-4B-Q4_K_M.gguf + +- !!merge <<: *spark-x2-5-4b + name: "spark-x2.5-4b-q8" + variants: [] description: | - # DeepSeek-V4-Flash-Vision-Exp - - ## Introduction - - We are excited to introduce **DeepSeek-V4-Flash-Vision-Exp**, our first experimental multimodal model in the DeepSeek-V4 family. It builds on the DeepSeek-V4-Flash architecture by incorporating visual modules and undergoing continued training to unlock visual understanding capabilities. - - Compared to DeepSeek-V4-Flash-0731, DeepSeek-V4-Flash-Vision-Exp achieves substantial improvements on its multimodal agent capabilities, while maintaining comparable performance on text-only agent tasks. - - Notes: - - 1. For the text agent benchmarks above, DeepSeek models are evaluated with the minimal mode of DeepSeek Harness as the agent framework, using the `max` reasoning effort level with `temperature = 1.0, top_p = 0.95`. - 2. † For ApexBench and Agents' Last Exam, DeepSeek-V4-Flash-0731 ignores the multimodal elements in the input. - - ## Repository layout - - This repository contains the tokenizer, prompt encoding reference, and a - minimal PyTorch inference implementation for DeepSeek-V4 Flash Vision. The - reference inference covers the vision encoder and aligner, DFlash attention, - MoE, Hyper-Connections, and the DSpark forward path. - - ... - license: "mit" - tags: - - llm - - gguf - - deepseek - icon: https://github.com/deepseek-ai/DeepSeek-V2/blob/main/figures/logo.svg + Spark-X2.5-4B is XHToken's 4B text model for conversation, reasoning, + coding, and multilingual tasks. This build uses Q8_0 GGUF weights, + the embedded Jinja chat template, and a 32K-token default context. overrides: - backend: ds4 + backend: llama-cpp + context_size: 32768 function: + automatic_tool_parsing_fallback: true grammar: disable: true known_usecases: - chat + options: + - use_jinja:true parameters: - model: ds4flash.gguf + model: llama-cpp/models/Spark-X2.5-4B-Q8_0/Spark-X2.5-4B-Q8_0.gguf + temperature: 1 + top_p: 0.95 + top_k: -1 + min_p: 0 + repeat_penalty: 1 template: use_tokenizer_template: true files: - - filename: ds4flash.gguf - sha256: "" - uri: https://huggingface.co/unsloth/DeepSeek-V4-Flash-Vision-Exp-GGUF + - filename: llama-cpp/models/Spark-X2.5-4B-Q8_0/Spark-X2.5-4B-Q8_0.gguf + sha256: 5c2c3c190e4337e1016b8593ca8e26e8b18c972200b107385d4ec61a25d9dea2 + uri: https://huggingface.co/XHToken/Spark-X2.5-4B-GGUF/resolve/d313463f1ff7dc29db20193fddc4ad8386d16261/Spark-X2.5-4B-Q8_0.gguf + - name: "qwopus3.8-27b-flash" variants: - model: qwopus3.8-27b-flash-q8