Crate nxpu

Structs§

apply_dynamic_batch 🔒: Apply --dynamic-batch: for every storage-buffer global variable whose type is an Array (the common WGSL pattern), wrap it in a rank-2+ Tensor type with the first dimension set to Symbolic("batch").
build_registry 🔒
load_constant_tensors 🔒: Read <tensor-name>.bin for every tensor the caller supplied.
main 🔒
parse_calibration_method 🔒
parse_opt_level 🔒
parse_precision 🔒
run 🔒
write_output_file 🔒