Expand description
ORT execution-provider configuration and session loading.
Structsยง
- GpuSession
Config ๐ - Configuration for loading an ORT session and tokenizer for a single GPU worker.
Functionsยง
- execution_
providers ๐ - Returns the execution providers to use for this platform and EP selection.
- load_
models ๐ - Downloads (if not already cached) and loads both the ORT session and the tokenizer for the given model variant, returning them as a pair.
- load_
session ๐ - Builds an ORT session from the ONNX model file with the given execution providers.