huankimtran

DeepSeek R1 Distill: Complete Tutorial for Deployment & Fine-Tuning

This guide shows how to deploy an uncensored DeepSeek R1 Distill model to Google Cloud Run with GPU support and how to perform a basic, functional fine-tuning process. The tutorial is split into:

1. Environment Setup
2. FastAPI Inference Server
3. Docker Configuration
4. Google Cloud Run Deployment
5. Fine-Tuning Pipeline (Cold Start, Reasoning RL, Data Collection, Final RL Phase)

huankimtran

Intro

TCL-Expect scripts are an amazingly easy way to script out laborious tasks in the shell when you need to be interactive with the console. Think of them as a "macro" or way to programmaticly step through a process you would run by hand. They are similar to shell scripts but utilize the .tcl extension and a different #! call.

Setup Your Script

The first step, similar to writing a bash script, is to tell the script what it's executing under. For expect we use the following:

#!/usr/bin/expect