Author Uses HuggingChat's ML Intern to Build Multiple Small Models on a Budget
Understand the story
According to a first-hand report on the Hugging Face Blog dated 2026-10-08, an author—because only a compressed version of the official Qwen-Image 2.1 9B prompt rewriter was available on the Hugging Face Hub, with no smaller runnable version—used HuggingChat's ML-intern mode to create 7 models over two days at a total cost of about US$103, including a 0.8B student model that can run on CPU. The information disclosed so far comes mainly from the author's blog, which shows the complete process and itemized costs of batch distilling/building small models at low cost using the ML-intern mode.
AI overview of reporting · Updated 9 days ago
Reporting timeline
Explore different aspects of the story through its reporting.
- Hugging Face BlogSelectedThe model that didn't exist, so you made it yourself
Because the Hub only had a compressed version of the official 9B prompt rewriter for Qwen-Image 2.1, the author used HuggingChat's ML-intern mode to build a 0.8B student version that runs on CPU.