mudler · localai-bot · Nov 4, 2025
diff --git a/gallery/index.yaml b/gallery/index.yaml
@@ -3295,7 +3295,7 @@
    - gemma3
    - gemma-3
  overrides:
    #mmproj: gemma-3-27b-it-mmproj-f16.gguf
    parameters:
      model: gemma-3-27b-it-Q4_K_M.gguf
  files:
@@ -3313,7 +3313,7 @@
  description: |
    google/gemma-3-12b-it is an open-source, state-of-the-art, lightweight, multimodal model built from the same research and technology used to create the Gemini models. It is capable of handling text and image input and generating text output. It has a large context window of 128K tokens and supports over 140 languages. The 12B variant has been fine-tuned using the instruction-tuning approach. Gemma 3 models are suitable for a variety of text generation and image understanding tasks, including question answering, summarization, and reasoning. Their relatively small size makes them deployable in environments with limited resources such as laptops, desktops, or your own cloud infrastructure.
  overrides:
    #mmproj: gemma-3-12b-it-mmproj-f16.gguf
    parameters:
      model: gemma-3-12b-it-Q4_K_M.gguf
  files:
@@ -3331,7 +3331,7 @@
  description: |
    Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. Gemma 3 models are multimodal, handling text and image input and generating text output, with open weights for both pre-trained variants and instruction-tuned variants. Gemma 3 has a large, 128K context window, multilingual support in over 140 languages, and is available in more sizes than previous versions. Gemma 3 models are well-suited for a variety of text generation and image understanding tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Gemma-3-4b-it is a 4 billion parameter model.
  overrides:
    #mmproj: gemma-3-4b-it-mmproj-f16.gguf
    parameters:
      model: gemma-3-4b-it-Q4_K_M.gguf
  files:
@@ -6949,7 +6949,7 @@
      sha256: 2756551de7d8ff7093c2c5eec1cd00f1868bc128433af53f5a8d434091d4eb5a
      uri: huggingface://Triangle104/Nano_Imp_1B-Q8_0-GGUF/nano_imp_1b-q8_0.gguf
 - &qwen25
  name: "qwen2.5-14b-instruct" ## Qwen2.5
  icon: https://avatars.githubusercontent.com/u/141221163
  url: "github:mudler/LocalAI/gallery/chatml.yaml@master"
  license: apache-2.0
@@ -23049,3 +23049,83 @@
     - filename: YanoljaNEXT-Rosetta-27B-2511.i1-Q4_K_M.gguf
       sha256: 0a599099e93ad521045e17d82365a73c1738fff0603d6cb2c9557e96fbc907cb
       uri: huggingface://mradermacher/YanoljaNEXT-Rosetta-27B-2511-i1-GGUF/YanoljaNEXT-Rosetta-27B-2511.i1-Q4_K_M.gguf
+- !!merge <<: *llama3
+  name: "lightonocr-1b-1025"
+  urls:
+    - https://huggingface.co/noctrex/LightOnOCR-1B-1025-GGUF
+  description: |
+    **Model Name:** LightOnOCR-1B-1025
+    **Repository:** [lightonai/LightOnOCR-1B-1025](https://huggingface.co/lightonai/LightOnOCR-1B-1025)
+    **License:** Apache 2.0
+    **Pipeline:** Image-to-Text (OCR & Document Understanding)
+    **Languages:** English, French, German, Spanish, Italian, Dutch, Portuguese, Swedish, Danish
+
+    ---
+
+    ### 🔍 **Description**
+
+    LightOnOCR-1B-1025 is a compact, end-to-end vision-language model designed for high-accuracy Optical Character Recognition (OCR) and document understanding. Built on a Pixtral-based vision encoder and a Qwen3-derived text decoder, it delivers state-of-the-art performance in its size category while being significantly faster and more cost-effective than larger general-purpose models.
+
+    This model excels at extracting structured text from complex documents—handling tables, forms, receipts, multi-column layouts, and mathematical notation—without relying on external OCR pipelines.
+
+    ---
+
+    ### ⚡ **Key Features**
+
+    - **Speed:** Up to 5× faster than dots.ocr, 2× faster than PaddleOCR-VL-0.9B
+    - **Efficiency:** Processes ~5.71 pages per second on a single H100 (~493k pages/day) at under $0.01 per 1,000 pages
+    - **Multilingual Support:** Trained on diverse multilingual PDFs (Latin script)
+    - **End-to-End Architecture:** Fully differentiable; ideal for fine-tuning and integration
+    - **Optimized for Real-World Use:** Works well with PDFs rendered at ~1540px longest edge
+
+    ---
+
+    ### 📊 **Performance Highlights (Olmo-Bench)**
+
+    | Task             | Score |
+    |------------------|-------|
+    | Overall Accuracy | **76.1** |
+    | Multi-Column     | 80.0 |
+    | Tables           | 35.2 |
+    | Tiny Text        | 88.7 |
+
+    ---
+
+    ### 🧩 **Use Cases**
+
+    - Automated document processing
+    - Receipt and invoice parsing
+    - Scientific paper and book OCR
+    - Form and table extraction
+    - Low-cost, scalable OCR for enterprise workflows
+
+    ---
+
+    ### 📦 **Variants Available**
+
+    - **`LightOnOCR-1B-1025` (default)** – Full multilingual model (151k vocab)
+    - **`LightOnOCR-1B-32k`** – Fast, pruned vocabulary (32k tokens), optimized for European languages
+    - **`LightOnOCR-1B-16k`** – Most compact variant (16k tokens), smallest memory footprint
+
+    ---
+
+    ### 🚀 **Getting Started**
+
+    Run with vLLM for blazing-fast inference:
+
+    ```bash
+    vllm serve lightonai/LightOnOCR-1B-1025 --limit-mm-per-prompt '{"image": 1}' --async-scheduling
+    ```
+
+    👉 **[Try the demo](https://huggingface.co/spaces/lightonai/LightOnOCR-1B-Demo)** | 📝 **[Read the blog](https://huggingface.co/blog/lightonai/lightonocr/)**
+
+    ---
+
+    **Ideal for developers, researchers, and enterprises seeking fast, accurate, and affordable document intelligence.**
+  overrides:
+    parameters:
+      model: LightOnOCR-1B-1025-Q4_K_M.gguf
+  files:
+    - filename: LightOnOCR-1B-1025-Q4_K_M.gguf
+      sha256: da36fb008a81128553933a15dc6373c1d0692e3ed1c17e9115521d84c473dbd5
+      uri: huggingface://noctrex/LightOnOCR-1B-1025-GGUF/LightOnOCR-1B-1025-Q4_K_M.gguf