How I verify an ONNX model swap by hashing tensors (and what it still doesn't prove)
We run background removal as an ONNX model inside a container behind a Cloudflare Worker. Last week we split one combined runtime image into two profile images — a "Lite" profile and a "Lite-2K" profile — and shrunk the distributed weights. The interesting part was not the image split. It was provi…
Read the full story at DEV Community — Machine Learning ↗
Timeline · 1 report
- 2026-11-12 16:23 · DEV Community — Machine Learning
How I verify an ONNX model swap by hashing tensors (and what it still doesn't prove)