Upload README.md with huggingface_hub

README.md

@@ -1,14 +1,17 @@
-# Hexo-0.5-Mini
+---
+base_model: WeiboAI/VibeThinker-1.5B
+library_name: peft
+tags:
+- llama-cpp
+- gguf
+- loRA
+---
 
-- Author: `LegoKeeper`
-- Base model: `WeiboAI/VibeThinker-1.5B`
-- Fine-tuning: LoRA trained on Kaggle (project: Hexo_FT), then merged into the base model.
-- Quantization: `Q4_K_M` GGUF via `llama.cpp` (if present).
+# LegoKeeper/Hexo-0.5-Mini
 
-## Contents
+- **Base Model:** WeiboAI/VibeThinker-1.5B
+- **Quantization:** Q4_K_M
 
-- LoRA adapter + tokenizer files (uploaded from `hexo-vibethinker-lora-manual`) — for future fine-tuning or merging.
-- `Hexo-0.5-Mini-Q4_K_M.gguf` — merged and quantized GGUF suitable for use with:
-  - `llama.cpp`
-  - `ollama` (after import)
-  - any GGUF-compatible runtime.
+## Usage
+
+This repo contains the merged FP16 model (in safetensors), the LoRA adapter, and the GGUF file `Hexo-0.5-Mini-Q4_K_M.gguf`.