A model artifact is the deliverable you actually deploy: a file containing the learned weights plus a description of the network. Formats such as ONNX, GGUF, or a framework checkpoint determine which hardware and serving tools can load it, so the format is a procurement decision, not a detail.
The book · Volume 6 of 7
AI for Video: A Complete Guide to Intelligence in Video Products: From Models and Inference to Computer Vision, Speech, Search, and Generative Video
The model artifact format decides which runtimes and hardware can serve a model, so it is a procurement choice as much as a technical one. This volume covers model artifacts in depth: ONNX, GGUF and framework checkpoints, quantization, the open-versus-closed procurement decision, and deployment topologies. Written by Nikolay Sapunov, CEO at Fora Soft.
Unsure which model artifact format fits your serving stack?
Tell us your model and target hardware, and we will recommend an artifact format and conversion path with measured speed on your runtime. 250+ video products shipped since 2005.
