Posts by Arttu Niemela
Using Reinforcement Learning to Fix Text in AI-Generated Videos
- 25 November 2025
One common giveaway that a video is AI-generated is the text. Letters may look slightly malformed or nonsensical, words can be misspelled and full sentences can have grammatical errors. Improving text generation in videos isn’t just a cosmetic issue - it is essential to generate the prompted text precisely, lest the message become confusing, unprofessional, and potentially misleading. This is an excellent case for leveraging reinforcement learning to improve a video generation model on a specific task without requiring massive amounts of suitable training data.
Wan2.2 Fine-Tuning: Tailoring an Advanced Video Generation Model on a Single GPU
- 19 August 2025
This blog post will guide you through fine-tuning Wan2.2 - a state-of-the-art video generation model - on a single AMD Instinct MI300X GPU. By following this guide, you’ll unlock Wan2.2’s advanced video generation capabilities and customize the output — whether in a unique artistic style or a specialized domain — all while running memory efficiently even on a single GPU. Here are some examples of how you can put this guide into practice: