feat: Use Phi-4-mini-instruct

README.md

@@ -1,7 +1,7 @@
 ---
-title: Phi-4 Reasoning Demo
-emoji: 🤖
-colorFrom: purple
+title: Phi-4 Mini Instruct Demo
+emoji: 🧠
+colorFrom: blue
 colorTo: indigo
 sdk: gradio
 sdk_version: 5.29.0
@@ -9,20 +9,28 @@ app_file: app.py
 pinned: false
 ---
 
-# Phi-4 Reasoning Demo (Hugging Face Space)
+# Phi-4 Mini Instruct – Hugging Face Space
 
-This is a Gradio-powered demo using the `microsoft/phi-4-reasoning-plus` model from Hugging Face for advanced reasoning and text generation.
+This Space is a lightweight Gradio demo using the `microsoft/Phi-4-mini-instruct` model from Hugging Face for compact, instruction-tuned language generation.
 
 ## 🔍 Model
 
-- **Model:** [`microsoft/phi-4-reasoning-plus`](https://huggingface.co/microsoft/phi-4-reasoning-plus)
-- **Use Case:** Text generation, reasoning tasks, assistant-style prompts
+- **Model:** [`microsoft/Phi-4-mini-instruct`](https://huggingface.co/microsoft/Phi-4-mini-instruct)
+- **Type:** Small, instruction-tuned LLM
+- **Use Case:** Text generation for assistant-style tasks with low resource usage
 
 ## 🚀 Features
 
-- Simple Gradio interface
-- Accepts a user prompt and returns a generated response using the Phi-4 model
+- Lightweight and fast, great for CPU-only Spaces
+- Simple Gradio interface for trying out prompts
+- Hugging Face `transformers` integration
 
 ## 📦 Requirements
 
-Listed in `requirements.txt`:
+Installed via `requirements.txt`:
+
+```txt
+gradio>=5.29.0
+transformers>=4.51.3
+torch
+accelerate

app.py

@@ -2,7 +2,7 @@ import gradio as gr
 from transformers import pipeline
 
 
-generator = pipeline("text-generation", model="microsoft/phi-4-reasoning-plus", device_map="auto")
+generator = pipeline("text-generation", model="microsoft/microsoft/Phi-4-mini-instruct", device_map="auto")
 
 def chat_with_phi(prompt):
     response = generator(prompt, max_new_tokens=200, do_sample=False)[0]["generated_text"]