From ec4bd5119a267c617fe2b6eb8a23823b30b22412 Mon Sep 17 00:00:00 2001 From: aomukai Date: Sun, 16 Nov 2025 18:01:10 +0900 Subject: [PATCH] fix: Update model download links in installation instructions --- README.md | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/README.md b/README.md index 92b0d38..fded7c3 100644 --- a/README.md +++ b/README.md @@ -74,8 +74,7 @@ Unlike its predecessor, this new version is built with **JavaScript + HTML**, ma - [CPU only](https://github.com/ggerganov/llama.cpp/releases) - Download `llama-*-bin-win-avx2-x64.zip` - Extract `llama-server.exe` to the Writingway folder 4. **Download a model** (optional, for local AI): - - [Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct-GGUF) (2.5GB, fast) - - [Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-GGUF) (5GB, better quality) + - [Qwen3-4B-Instruct](https://huggingface.co/unsloth/Qwen3-4B-Thinking-2507-GGUF/resolve/main/Qwen3-4B-Thinking-2507-Q4_K_M.gguf?download=true) - Put the `.gguf` file in the `models/` folder 5. Double-click **`start.bat`** 6. Writingway opens at **http://localhost:8000**