{

"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"colab": {
"provenance": [],
"gpuType": "T4"
},
"kernelspec": {
"name": "python3",
"display_name": "Python 3"
},
"language_info": {
"name": "python"
},
"accelerator": "GPU"
},
"cells": [
{
"cell_type": "code",
"source": [
"# Model selection\n",
"MODEL_NAME = \"maryasov/qwen2.5-coder-cline:7b-instruct-q8_0\"\n",
"%env OLLAMA_CONTEXT_LENGTH=16384\n",
"%env OLLAMA_HOST=0.0.0.0\n",
"%env OLLAMA_KEEP_ALIVE=-1\n"
],
"metadata": {
"id": "WG0UDdk86dxb"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!apt-get install -y lshw pciutils\n",
"!nvcc --version\n",
"!nvidia-smi\n",
"\n",
"from psutil import virtual_memory\n",
"ram_gb = virtual_memory().total / 1e9\n",
"print(f\"\\n🧠 Available RAM: {ram_gb:.1f} GB\")\n",
"print(\"✅ High-RAM runtime!\" if ram_gb >= 20 else \"❌ Not a high-RAM
runtime.\")\n"
],
"metadata": {
"id": "VskPhA1M6h8j"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!curl -fsSL https://ollama.com/install.sh | sh\n"
],
"metadata": {
"id": "J83WxdLL6k75"
 },
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"import subprocess\n",
"import time\n",
"import requests\n",
"import threading\n",
"\n",
"# Start ollama serve in a background thread\n",
"def start_ollama():\n",
" subprocess.call(['ollama', 'serve'])\n",
"\n",
"ollama_thread = threading.Thread(target=start_ollama)\n",
"ollama_thread.daemon = True\n",
"ollama_thread.start()\n",
"\n",
"# Pull model (this also verifies Ollama CLI is ready)\n",
"!ollama pull {MODEL_NAME}\n",
"\n",
"# Wait for Ollama HTTP API to be ready\n",
"def wait_for_ollama(timeout=60):\n",
" for i in range(timeout):\n",
" try:\n",
" r = requests.get(\"http://localhost:11434\")\n",
" if r.status_code in [200, 404]:\n",
" print(f\"✅ Ollama is up (after {i+1}s).\")\n",
" return\n",
" except requests.exceptions.ConnectionError:\n",
" pass\n",
" print(f\"⏳ Waiting for Ollama to start... {i+1}s\")\n",
" time.sleep(1)\n",
" raise RuntimeError(\"❌ Ollama did not start in time.\")\n",
"\n",
"wait_for_ollama()\n"
],
"metadata": {
"id": "4uR5FDlu6nav"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"!wget https://github.com/cloudflare/cloudflared/releases/latest/download/
cloudflared-linux-amd64 -O cloudflared\n",
"!chmod +x cloudflared\n"
],
"metadata": {
"id": "_XUWhodx6pTh"
},
"execution_count": null,
"outputs": []
},
{
 "cell_type": "code",
"source": [
"import re\n",
"\n",
"# Run cloudflared tunnel in background and get the public URL\n",
"cloudflared_proc = subprocess.Popen(\n",
" ['./cloudflared', 'tunnel', '--url', 'http://localhost:11434', '--no-
autoupdate'],\n",
" stdout=subprocess.PIPE,\n",
" stderr=subprocess.STDOUT,\n",
" text=True\n",
")\n",
"\n",
"public_url = None\n",
"for line in cloudflared_proc.stdout:\n",
" print(line.strip())\n",
" match = re.search(r'(https://.*\\.trycloudflare\\.com)', line)\n",
" if match:\n",
" public_url = match.group(1)\n",
" break\n",
"\n",
"if public_url:\n",
" print(f\"\\n✅ Public URL for Ollama:\\n{public_url}\")\n",
"else:\n",
" raise RuntimeError(\"❌ Could not find public Cloudflare URL.\")\n"
],
"metadata": {
"id": "eVmbIF6b6qsk"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"import json\n",
"\n",
"data = {\n",
" \"model\": MODEL_NAME,\n",
" \"prompt\": \"Question: What is the capital of Japan?\\nAnswer:\",\n",
" \"stream\": False\n",
"}\n",
"\n",
"response = requests.post(f\"{public_url}/api/generate\", json=data)\n",
"print(response.json())\n"
],
"metadata": {
"id": "b4oAfEAG6sfb"
},
"execution_count": null,
"outputs": []
}
]
}