Spaces:
Sleeping
Sleeping
| title: Ml Sensei Chat | |
| emoji: 📈 | |
| colorFrom: red | |
| colorTo: yellow | |
| sdk: gradio | |
| sdk_version: 6.0.2 | |
| app_file: app.py | |
| pinned: false | |
| license: mit | |
| short_description: A place to chat to an ML training LoRA of TinyLlama. | |
| # 🥋 ML Sensei — TinyLlama LoRA Chat | |
| *A lightweight interactive tutor for machine learning concepts.* | |
| Welcome to **ML Sensei**, a fine-tuned TinyLlama model designed to **teach machine learning concepts with clarity, intuition, and gentleness**. | |
| This Space provides an interactive chat interface where you can ask ML Sensei questions about: | |
| - Machine Learning | |
| - Deep Learning | |
| - Neural networks | |
| - Optimization | |
| - Training loops | |
| - Data science fundamentals | |
| - Applied concepts and intuition | |
| ML Sensei uses a **LoRA adapter** fine-tuned on a custom instruction dataset built from curated ML texts, chunk-based self-generated Q&A, intuitive breakdowns, and code-ready explanations. | |
| --- | |
| ## 🧠 Model Details | |
| **Base Model:** | |
| [TinyLlama/TinyLlama-1.1B-Chat-v1.0](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0) | |
| **LoRA Adapter:** | |
| mackenzietechdocs/ml-sensei-lora-tinyllama-1.1b | |
| **Architecture:** | |
| - Decoder-only Transformer | |
| - Chat-tuned | |
| - LoRA-adapted for ML instruction | |
| --- | |
| ## 🧩 What ML Sensei Can Do | |
| ML Sensei specializes in: | |
| ### ✔ Beginner-friendly explanations | |
| Breaks down ML ideas into approachable, intuitive pieces. | |
| ### ✔ Simple analogies + concept-first teaching | |
| Uses everyday intuition instead of heavy math when possible. | |
| ### ✔ Helpful follow-up questions | |
| Ask more and ML Sensei will guide you deeper step-by-step. | |
| ### ✔ Code examples on request | |
| Provides PyTorch or pseudocode snippets when helpful. | |
| ### ✔ Checking your understanding | |
| Ask ML Sensei to quiz you on a topic! | |
| --- | |
| ## ▶️ How to Use This Space | |
| Just type a question into the chat box: | |
| Example prompts: | |
| - *“Explain gradient descent simply.”* | |
| - *“What is weight decay?”* | |
| - *“How does cross-entropy work conceptually?”* | |
| - *“Can you quiz me on neural networks?”* | |
| - *“Write a small PyTorch example showing a training loop.”* | |
| ML Sensei will respond with high-quality educational text aimed at clarity and intuition. | |
| --- | |
| ## ⚙️ Technical Details | |
| This app loads: | |
| 1. The **TinyLlama 1.1B Chat model** | |
| 2. The **ML Sensei LoRA adapter** via PEFT | |
| 3. A **Gradio chat interface** for interactive conversation | |
| 4. TinyLlama’s official chat template for structured inputs | |
| The LoRA adds: | |
| - Improved conceptual explanations | |
| - Beginner-friendly language | |
| - Educational reasoning patterns | |
| - Stepwise breakdowns | |
| - Vocabulary simplification | |
| The underlying LLM remains safe, stable, and compact enough to run on **CPU**. | |
| --- | |
| ## 📦 Environment & Requirements | |
| The Space installs: | |
| ``` | |
| transformers | |
| peft | |
| gradio | |
| accelerate | |
| torch | |
| sentencepiece | |
| ``` | |
| Inference runs on CPU unless upgraded to GPU. | |
| --- | |
| ## 📜 License | |
| - Base model license: TinyLlama License | |
| - LoRA adapter license: Apache 2.0 | |
| - Dataset used: Fully model-generated instructional text | |
| Users must comply with the base model license. | |
| --- | |
| ## 🛡 Safety | |
| ML Sensei is designed for **educational purposes only**. | |
| It is **not** a certified resource for: | |
| - Academic evaluation | |
| - Recruitment / hiring | |
| - Safety-critical decision making | |
| Please report issues via “Community” tab. | |
| --- | |
| ## 🤝 Contributing | |
| If you'd like to extend ML Sensei with: | |
| - More curriculum topics | |
| - More examples | |
| - Better reasoning chains | |
| - Additional Q&A datasets | |
| Feel free to open a PR or connect with me on GitHub or Hugging Face! | |
| --- | |
| ## ✨ Philosophy | |
| > **Machine learning becomes easier when taught patiently and simply.** | |
| ML Sensei exists to make ML concepts less intimidating — and to help more learners enter the AI field confidently. | |
| Enjoy chatting with your tutor, and may your gradients always descend! 🧘♀️📉 | |