Run Ollama with Qwen in Docker Space

Dockerfile

@@ -0,0 +1,31 @@
+FROM ubuntu:22.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    curl \
+    ca-certificates \
+    python3 \
+    python3-pip \
+    bash \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install Ollama
+RUN curl -fsSL https://ollama.com/install.sh | sh
+
+WORKDIR /app
+
+COPY requirements.txt /app/requirements.txt
+RUN pip3 install --no-cache-dir -r /app/requirements.txt
+
+COPY app.py /app/app.py
+COPY start.sh /app/start.sh
+RUN chmod +x /app/start.sh
+
+# Hugging Face Docker Spaces must listen on 7860
+EXPOSE 7860
+
+# Change this to a different Qwen model if needed
+ENV OLLAMA_MODEL=qwen2.5:0.5b
+
+CMD ["/app/start.sh"]

README.md

@@ -1,10 +1,19 @@
 ---
-title: Llm
-emoji: 🐠
-colorFrom: red
-colorTo: blue
+title: Ollama Qwen Demo
+emoji: 🤖
+colorFrom: blue
+colorTo: green
 sdk: docker
-pinned: false
+app_port: 7860
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+This Space runs a local Ollama model with a Gradio UI.
+
+Default model:
+- `qwen2.5:0.5b`
+
+You can change model by editing `OLLAMA_MODEL` in `Dockerfile`.
+
+Notes:
+- Free CPU Spaces are limited, so smaller Qwen variants are recommended.
+- Larger models may fail or take a long time to pull.

app.py

@@ -0,0 +1,38 @@
+import os
+import requests
+import gradio as gr
+
+OLLAMA_URL = os.getenv("OLLAMA_URL", "http://127.0.0.1:11434")
+MODEL = os.getenv("OLLAMA_MODEL", "qwen2.5:0.5b")
+APP_KEY = os.getenv("APP_KEY", "")
+
+
+def chat(prompt: str, key: str):
+    if APP_KEY and key != APP_KEY:
+        return "Unauthorized"
+
+    try:
+        res = requests.post(
+            f"{OLLAMA_URL}/api/generate",
+            json={"model": MODEL, "prompt": prompt, "stream": False},
+            timeout=120,
+        )
+        res.raise_for_status()
+        return res.json().get("response", "No response from model")
+    except Exception as e:
+        return f"Error: {e}"
+
+
+with gr.Blocks(title="Ollama Qwen Demo") as demo:
+    gr.Markdown("# Ollama Qwen Demo")
+    gr.Markdown(f"Using model: `{MODEL}`")
+
+    prompt = gr.Textbox(label="Prompt", lines=6, placeholder="Ask Qwen anything...")
+    key = gr.Textbox(label="Access key (optional)", type="password")
+    output = gr.Textbox(label="Response", lines=12)
+
+    submit = gr.Button("Generate")
+    submit.click(chat, inputs=[prompt, key], outputs=output)
+
+
+demo.launch(server_name="0.0.0.0", server_port=7860)

requirements.txt

@@ -0,0 +1,2 @@
+gradio==4.44.1
+requests==2.32.3

start.sh

@@ -0,0 +1,21 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# Start Ollama in the background
+ollama serve &
+
+# Wait for Ollama API to become ready
+for i in {1..60}; do
+  if curl -sf http://127.0.0.1:11434/api/tags >/dev/null; then
+    break
+  fi
+  sleep 1
+done
+
+# Pull model if missing
+if ! ollama list | awk '{print $1}' | grep -qx "${OLLAMA_MODEL}"; then
+  echo "Pulling model ${OLLAMA_MODEL}..."
+  ollama pull "${OLLAMA_MODEL}"
+fi
+
+exec python3 app.py