Titaniumclock/Gemma-4-Rust-Coder

1---2tags:3- gguf4- llama.cpp5- unsloth6- vision-language-model7- rust8- coding9license: mit10datasets:11- Fortytwo-Network/Strandset-Rust-v112base_model:13- google/gemma-4-E4B-it14---15 16# Gemma-4-Rust-Coder : GGUF17 18This model is a specialized fine-tune of **Gemma 4**, specifically optimized for **Rust systems programming**, memory safety patterns, and high-performance development. It was trained using **Unsloth Studio** to ensure maximum efficiency and performance.19 20## 🦀 Fine-Tuning Focus21The model has been adjusted to excel in:22* **Idiomatic Rust:** Writing clean, "Rusty" code using modern patterns.23* **Concurrency:** Deep understanding of `Send`, `Sync`, and async runtimes like `Tokio`.24* **Vision-to-Code:** Using its multimodal capabilities to translate architecture diagrams or UI mockups into functional Rust code.25 26## 🤝 Credits & Acknowledgments27Special thanks to **Fortytwo-Network** for providing the **[Strandset-Rust-v1](https://huggingface.co/datasets/Fortytwo-Network/Strandset-Rust-v1)** dataset. This model's specialized knowledge of the Rust ecosystem is a direct result of this high-quality data.28 29## 🚀 Usage30This model is converted to GGUF format for seamless use with `llama.cpp` and other compatible executors.31 32**Example usage**:33* **Text-only LLM:** `llama-cli -hf MassivDash/Gemma-4-Rust-Coder --jinja`34* **Multimodal / Vision:** `llama-mtmd-cli -hf MassivDash/Gemma-4-Rust-Coder --jinja`35 36## 📂 Available Model files:37* `gemma-4-e2b-it.Q3_K_M.gguf`38* `gemma-4-e2b-it.BF16-mmproj.gguf`39 40## ⚠️ Ollama Note for Vision Models41**Important:** Ollama currently does not support separate `mmproj` files for vision models.42 43To create an Ollama model from this vision model:441. Place the `Modelfile` in the same directory as the finetuned bf16 merged model.452. Run: `ollama create model_name -f ./Modelfile`46   *(Replace `model_name` with your desired name)*47 48## 🔗 Stay Connected49For more insights on AI development and fine-tuning, visit my blog: 50👉 **[spaceout.pl](https://spaceout.pl)**51 52---53 54*This model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth)*55 56[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)