Qwen 3.8 27B — Multi-Kaggle Cluster

Authenticated as admin · Native CUDA C++ Dual-T4 Engine · 64K Ultra Context

Active GPU Nodes

- / -

Checking nodes...

Total Tokens Processed

0 Prompt · 0 Output

Avg Cluster Speed

0.0 tok/s

MTP Speculative Decoding Ready

Total Requests

0 ms avg latency

Live Kaggle Nodes & GPU Quota Control

ROUND_ROBIN

Node / Account	Hardware & Flags	Live Speeds	GPU Quota (30h/wk)	Status	Last Heartbeat	Actions
Loading node telemetry...

Load Balancer & Routing Policy

Strategy

Real-Time Generation Speed (tok/s)

End-to-End Latency (ms)

Token Volume per Request (Recent)

Client API Keys (for IDE / Cursor / Apps)

Key Name	Created	Requests	Total Tokens	Last Used	Status	Actions

Kaggle Account Roster & Weekly Quotas

Account ID	Display Name	Kaggle User	Weekly Limit	Roster Pool	Live Kaggle GPU	Actions
Loading accounts...

Paginated Traffic Logs

Timestamp	Served By	Method	Endpoint	Tokens (In / Out)	Speed	Status	Latency

Showing 0-0 of 0 requests

Page 1 of 1

Cluster Credentials & Endpoints

OpenAI API Base URL

https://qwen-proxy.gauravsuthar980.workers.dev/v1

Gateway Worker URL

https://qwen-proxy.gauravsuthar980.workers.dev

Shared Admin Secret

qwen-secret-a87f2e19b4c036d7e98a2f5c1d4b6e8a

Model Identifier

qwen3.8-27b 64K Context

Reasoning aliases: qwen3.8-27b-fast, qwen3.8-27b-high

Step 1: Configure Kaggle Notebook Settings & Secrets

In your Kaggle notebook, verify the right sidebar settings:

Accelerator: GPU T4 x2 (Dual T4) Internet: On (Required) Persistence: No persistence

Go to Settings → Add-ons → Secrets in your notebook and add the following keys:

Secret Key	Recommended Value	Description
`WORKER_URL`	https://qwen-proxy.gauravsuthar980.workers.dev	Your Cloudflare Worker Gateway URL
`ADMIN_TOKEN`	qwen-secret-a87f2e19b4c036d7e98a2f5c1d4b6e8a	Admin token for tunnel registration & watchdog
`ACCOUNT_ID`	gaurav (or `rahul`, `friend-2`)	Unique node ID used for load balancing
`ACCOUNT_NAME`	Gaurav Dual-T4	Display name shown on the live cluster table

Step 2: Serving Script (02_serve_qwen_llamacpp.py)

Paste this script into your Kaggle Notebook cell. It automatically compiles native CUDA llama-server with Flash Attention & MTP Speculative Decoding, creates a Cloudflare tunnel, and connects to this control plane.

02_serve_qwen_llamacpp.py — Native C++ Dual-T4 Server

Step 3: Connect Applications (IDE, Cursor, Python, cURL)

Python (OpenAI SDK 1.0+) with Streaming Reasoning

cURL Streaming Request

Cursor / Continue.dev / Cline Custom Model Configuration

LangChain ChatOpenAI Setup

Step 4: Multi-Account Cluster Setup (Pool 60h – 120h+ Free GPU Quota)

Kaggle grants 30 hours of Dual-T4 GPU per account per week. By connecting 2 or more accounts, your cluster automatically pools their quotas (2 accounts = 60 hrs/week, 3 accounts = 90 hrs/week, 4 accounts = 120 hrs/week) with automatic load balancing and failover!

1. Register in Roster

Go to API Keys & Roster tab and add Account ID (e.g. account2), Display Name (e.g. Rahul Dual-T4), and Kaggle Username.

2. Add Kaggle Secrets in Account 2

In Account 2's notebook, add WORKER_URL, ADMIN_TOKEN, and set ACCOUNT_ID to account2. Attach GGUF and click Run All.

3. Smart Quota Routing

Under Overview & Nodes, select GPU-Hour-Aware (Smart Quota) to automatically route prompts to whichever account has the most remaining free quota!

🚀 Optional: Remote 1-Click Start for Account 2

In your GitHub Repo (Settings → Secrets and variables → Actions), add KAGGLE_USERNAME_2 and KAGGLE_KEY_2. Now clicking Start Node for Account 2 in your dashboard will automatically launch the notebook via GitHub Actions!

Cluster Admin Sign In