From c4273485d19fc1b1063aa415b13d10c322b28e1f Mon Sep 17 00:00:00 2001 From: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com> Date: Wed, 7 Oct 2026 00:04:08 +0000 Subject: [PATCH] chore(gallery): add GUI-Owl 1.5 2B variants Add Q4_K_M and Q8_0 GGUF builds for screenshot interpretation and GUI planning with llama.cpp. Include the F16 vision projector in both builds and document installation and the separate action executor requirement. Pin all downloads and verify SHA256 values against Hugging Face LFS metadata and download headers. Assisted-by: Codex:gpt-6 --- docs/content/features/model-gallery.md | 23 +++++++ gallery/index.yaml | 88 ++++++++++++++++++++++++++ 2 files changed, 111 insertions(+) diff --git a/docs/content/features/model-gallery.md b/docs/content/features/model-gallery.md index 3bdfc6d41..e32d2554b 100644 --- a/docs/content/features/model-gallery.md +++ b/docs/content/features/model-gallery.md @@ -39,6 +39,29 @@ 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. +## GUI-Owl 1.5 2B Instruct + +GUI-Owl 1.5 uses Qwen3-VL to interpret screenshots and predict GUI actions. +The gallery includes Q4_K_M and Q8_0 GGUF builds for llama.cpp, each with an F16 vision projector. +Both use the embedded chat template and a 32,768-token context by default. +LocalAI serves the model; a separate agent must execute its predicted actions. + +Install with automatic variant selection: + +```bash +local-ai models install gui-owl-1.5-2b-instruct +``` + +To select Q8_0 explicitly: + +```bash +local-ai models install gui-owl-1.5-2b-instruct --variant gui-owl-1.5-2b-instruct-q8 +``` + +Send screenshots as image content in chat requests, as described in [GPT Vision](/features/gpt-vision/). +See the [model card](https://huggingface.co/mPLUG/GUI-Owl-1.5-2B-Instruct) +and [GGUF downloads](https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF). + ## Cyber-Ornith 1.5 9B Cyber-Ornith 1.5 is a Qwen3.5 fine-tune for security auditing, terminal tasks, and tool use. diff --git a/gallery/index.yaml b/gallery/index.yaml index 043e1542c..1820a24fb 100644 --- a/gallery/index.yaml +++ b/gallery/index.yaml @@ -22770,6 +22770,94 @@ - filename: mmproj/mmproj-Qwen3-VL-2B-Thinking-F16.gguf sha256: 4eabc90a52fe890d6ca1dad92548782eab6edc91f012a365fff95cf027ba529d uri: huggingface://unsloth/Qwen3-VL-2B-Thinking-GGUF/mmproj-F16.gguf +- name: gui-owl-1.5-2b-instruct + url: github:mudler/LocalAI/gallery/virtual.yaml@master + variants: + - model: gui-owl-1.5-2b-instruct-q8 + urls: + - https://huggingface.co/mPLUG/GUI-Owl-1.5-2B-Instruct + - https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF + description: | + GUI-Owl 1.5 is a Qwen3-VL-based model for interpreting screenshots and + planning GUI actions across desktop, mobile, and browser interfaces. + This 2B Instruct Q4_K_M GGUF build includes the F16 vision projector + and uses the embedded chat template with a 32K-token default context. + A separate agent must execute the predicted actions. + license: mit + tags: + - llm + - gguf + - cpu + - gpu + - vision + - multimodal + last_checked: "2026-10-07" + overrides: + backend: llama-cpp + context_size: 32768 + function: + grammar: + disable: true + known_usecases: + - chat + - vision + mmproj: GUI-Owl-1.5-2B-Instruct.mmproj-f16.gguf + options: + - use_jinja:true + parameters: + model: GUI-Owl-1.5-2B-Instruct.Q4_K_M.gguf + template: + use_tokenizer_template: true + files: + - filename: GUI-Owl-1.5-2B-Instruct.Q4_K_M.gguf + uri: https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF/resolve/a0592aa615540e507a918bc6281fdbfc1b441361/GUI-Owl-1.5-2B-Instruct.Q4_K_M.gguf + sha256: 5fd2c77781a14a91128928b5bd313252aab076b44682396af4bb2c99667f5ef5 + - filename: GUI-Owl-1.5-2B-Instruct.mmproj-f16.gguf + uri: https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF/resolve/a0592aa615540e507a918bc6281fdbfc1b441361/GUI-Owl-1.5-2B-Instruct.mmproj-f16.gguf + sha256: e3e3816edb92e62a51f8754b21f1369d8c0224f5488a878a4ef8284903bcf64a +- name: gui-owl-1.5-2b-instruct-q8 + url: github:mudler/LocalAI/gallery/virtual.yaml@master + urls: + - https://huggingface.co/mPLUG/GUI-Owl-1.5-2B-Instruct + - https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF + description: | + GUI-Owl 1.5 is a Qwen3-VL-based model for interpreting screenshots and + planning GUI actions across desktop, mobile, and browser interfaces. + This 2B Instruct Q8_0 GGUF build includes the F16 vision projector + and uses the embedded chat template with a 32K-token default context. + A separate agent must execute the predicted actions. + license: mit + tags: + - llm + - gguf + - cpu + - gpu + - vision + - multimodal + last_checked: "2026-10-07" + overrides: + backend: llama-cpp + context_size: 32768 + function: + grammar: + disable: true + known_usecases: + - chat + - vision + mmproj: GUI-Owl-1.5-2B-Instruct.mmproj-f16.gguf + options: + - use_jinja:true + parameters: + model: GUI-Owl-1.5-2B-Instruct.Q8_0.gguf + template: + use_tokenizer_template: true + files: + - filename: GUI-Owl-1.5-2B-Instruct.Q8_0.gguf + uri: https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF/resolve/a0592aa615540e507a918bc6281fdbfc1b441361/GUI-Owl-1.5-2B-Instruct.Q8_0.gguf + sha256: ed0901d94648da2e317dd577e3005914b8a721cb194dac1258557b214c19f5a4 + - filename: GUI-Owl-1.5-2B-Instruct.mmproj-f16.gguf + uri: https://huggingface.co/mradermacher/GUI-Owl-1.5-2B-Instruct-GGUF/resolve/a0592aa615540e507a918bc6281fdbfc1b441361/GUI-Owl-1.5-2B-Instruct.mmproj-f16.gguf + sha256: e3e3816edb92e62a51f8754b21f1369d8c0224f5488a878a4ef8284903bcf64a - name: qwen3-vl-2b-instruct url: github:mudler/LocalAI/gallery/qwen3.yaml@master urls: