grok

Dockerfile

@@ -1,36 +1,24 @@
-FROM python:3.11-slim
+FROM python:3.10-slim
 
-# 🔑 Critical: Export CMAKE_ARGS BEFORE installing deps
-ENV PYTHONDONTWRITEBYTECODE=1 \
-    PYTHONUNBUFFERED=1 \
-    PIP_NO_CACHE_DIR=1 \
-    UV_COMPILE_BYTECODE=1 \
-    PORT=7860 \
-    CMAKE_ARGS="-DLLAMA_CPU_ONLY=ON -DLLAMA_BLAS=OFF -DLLAMA_AVX=OFF"
+# Install system dependencies + Ollama
+RUN apt-get update && apt-get install -y curl && \
+    curl -fsSL https://ollama.ai/install.sh | sh && \
+    apt-get clean && rm -rf /var/lib/apt/lists/*
 
-# Minimal build tools
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    build-essential curl \
-    && rm -rf /var/lib/apt/lists/*
-
-# Install uv
-COPY --from=ghcr.io/astral-sh/uv:latest /uv /uv-bin/uv
-ENV PATH="/uv-bin:$PATH"
-
-# Non-root user (HF requirement)
-RUN useradd -m -u 1000 user && mkdir /app && chown -R user:user /app
-USER user
 WORKDIR /app
 
-# 🔥 Copy deps FIRST → better Docker cache
-COPY --chown=user pyproject.toml uv.lock ./
+# Copy requirements and install Python packages
+COPY requirements.txt .
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
 
-# ✅ Removed --no-build-isolation (breaks llama-cpp-python)
-RUN uv sync --frozen --no-dev
+# Copy all app files
+COPY . .
 
-# Copy ONLY your app code
-COPY --chown=user main.py ./
+# Make start script executable
+RUN chmod +x start.sh
 
+# Expose port (HF Spaces will use $PORT)
 EXPOSE 7860
-# ✅ No reload, bind to 0.0.0.0, use $PORT
-CMD ["uv", "run", "uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]
+
+# Start everything using the script
+CMD ["./start.sh"]

README.md

@@ -1,21 +1,23 @@
 ---
-title: Mann
-emoji: 🌍
-colorFrom: blue
-colorTo: yellow
+title: Mann - TinyLlama Chat
+emoji: 🧠
 sdk: docker
-pinned: false
 ---
 
-# Mann API
+# Mann - TinyLlama on Hugging Face Space
 
-FastAPI backend on HF Spaces.
+A simple FastAPI + Ollama deployment running **TinyLlama**.
 
-## Endpoints
+### Endpoints
 
-- `GET /` → status message
-- `GET /health` → `{"status":"ok"}`
+- `GET /` → Status
+- `POST /ask` → Normal response
+- `POST /stream` → Streaming response
 
-## Secrets
+### Example Request
 
-Add API keys in **Space Settings → Secrets** (never commit `.env`).
+```bash
+curl -X POST https://your-space.hf.space/ask \
+  -H "Content-Type: application/json" \
+  -d '{"text": "Hello, who are you?"}'
+```

app.py

@@ -0,0 +1,65 @@
+import os
+import logging
+from fastapi import FastAPI, HTTPException
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel
+from langchain_community.llms import Ollama
+from langchain.callbacks.manager import CallbackManager
+from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+app = FastAPI(title="TinyLlama on HF Space")
+
+MODEL_NAME = "tinyllama"
+PORT = os.getenv("PORT", 7860)
+
+def get_llm():
+    callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
+    return Ollama(
+        model=MODEL_NAME,
+        callback_manager=callback_manager,
+        temperature=0.7,
+        num_ctx=2048
+    )
+
+class Question(BaseModel):
+    text: str
+
+@app.get("/")
+def read_root():
+    return {
+        "message": f"✅ TinyLlama is running on HF Space!",
+        "model": MODEL_NAME,
+        "endpoints": ["/ask", "/stream"]
+    }
+
+@app.post("/ask")
+async def ask(question: Question):
+    try:
+        logger.info(f"Received question: {question.text[:100]}...")
+        llm = get_llm()
+        response = llm.invoke(question.text)
+        
+        return {
+            "question": question.text,
+            "answer": response.strip()
+        }
+    except Exception as e:
+        logger.error(f"Error: {e}")
+        raise HTTPException(status_code=500, detail=str(e))
+
+@app.post("/stream")
+async def stream(question: Question):
+    llm = get_llm()
+    
+    async def generate():
+        for chunk in llm.stream(question.text):
+            yield chunk
+    
+    return StreamingResponse(generate(), media_type="text/plain")
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=int(PORT))

requirements.txt

@@ -0,0 +1,5 @@
+fastapi
+uvicorn[standard]
+langchain
+langchain-community
+pydantic

start.sh

@@ -0,0 +1,35 @@
+#!/bin/bash
+
+echo "=== Starting Ollama + FastAPI on HF Space ==="
+
+# Optimization
+export OMP_NUM_THREADS=4
+export MKL_NUM_THREADS=4
+
+# Start Ollama server in background
+ollama serve &
+OLLAMA_PID=$!
+
+# Wait for Ollama to be ready
+echo "Waiting for Ollama to start..."
+for i in {1..40}; do
+    if curl -s http://localhost:11434/api/tags > /dev/null; then
+        echo "✅ Ollama is ready!"
+        break
+    fi
+    sleep 2
+done
+
+# Pull model if not present
+if ! ollama list | grep -q "tinyllama"; then
+    echo "Downloading tinyllama model... (first time only)"
+    ollama pull tinyllama
+else
+    echo "✅ tinyllama model already downloaded"
+fi
+
+echo "🚀 Starting FastAPI server on port ${PORT:-7860}"
+echo "API will be available at: http://0.0.0.0:${PORT:-7860}"
+
+# Start FastAPI
+uvicorn app:app --host 0.0.0.0 --port ${PORT:-7860} --workers 2