Quantized Model Checkpoint

Model assetModel ServingModelsarc:QuantizedModelCheckpoint

A framework-neutral model graph whose weights are stored at reduced integer or floating-point precision together with embedded quantization metadata (scaling factors, zero points), not yet compiled for a target GPU.

Responsibility. Holds the calibrated low-precision model and its quantization parameters as input for engine compilation.

Also known as: Quantized ONNX, INT8 ONNX model

Direct neighbourhood (hover for relationship types)

Relationships

is optimized by lifecycle

Engine Builder Ch7.4 Ref7.01

is produced by lifecycle

Model Quantizer Ch7.4 Ref7.01

Quantitative guidance

As stated by the sources; verify before use.

A 14 GB FP16 7B model becomes approximately 3.5 GB in INT8 (4x compression), speeding model loading and cutting storage cost (Ch7.4).

Classification

Technologies: NVIDIA TensorRT Model Optimizer (ModelOpt)ONNX
Quality attributes: Performance efficiency (ISO/IEC 25010)Maintainability (ISO/IEC 25010)

Sources

Ch7.4: T. Nguyen, "TensorRT-LLM Fundamentals and Quantization," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 7.4. ISBN: 9798244538229.
Ref7.01: NVIDIA, "Best practices," NVIDIA TensorRT Documentation. Accessed: Sep. 27, 2026. [Online]. Available: https://docs.nvidia.com/deeplearning/tensorrt/latest/performance/best-practices.html