| license: mit | |
| base_model: microsoft/Fara1.5-9B | |
| tags: | |
| - onnx | |
| - onnxruntime-genai | |
| - qwen3_5 | |
| - computer-use | |
| - multimodal | |
| - vision-language-model | |
| - text-generation | |
| # Fara1.5-9B ONNX (INT4) | |
| ONNX INT4 conversion of [microsoft/Fara1.5-9B](https://huggingface.co/microsoft/Fara1.5-9B) | |
| for use with [ONNX Runtime GenAI](https://github.com/microsoft/onnxruntime-genai) | |
| and the [ElBruno.LocalLLMs](https://github.com/elbruno/ElBruno.LocalLLMs) library. | |
| ## Model Description | |
| Fara1.5-9B is a multimodal computer use agent (CUA) fine-tuned from Qwen3.5-9B | |
| by Microsoft Research AI Frontiers. It observes browser screenshots and emits | |
| structured tool calls (click, type, scroll, navigate) for autonomous web tasks. | |
| ## Conversion Details | |
| | Field | Value | | |
| |---|---| | |
| | Source | microsoft/Fara1.5-9B | | |
| | Architecture | qwen3_5 / Qwen3VL (VisionGenAI) | | |
| | Precision | INT4 | | |
| | Context length | 32,768 tokens | | |
| | Builder | onnxruntime-genai model builder v0.14.1+ | | |
| ## Usage | |
| Use `LocalVisionChatClient` (Fara is a VisionGenAI / VLM model): | |
| ```csharp | |
| await using var client = new LocalVisionChatClient(new LocalLLMsOptions | |
| { | |
| Model = KnownModels.Fara15_9B, | |
| EnsureModelDownloaded = true | |
| }); | |
| ``` | |
| ## License | |
| MIT, inherited from the source model. | |