triton

Modules

modelopt.torch.quantization.triton.fp4_kernel	NVFP4 Fake Quantization Triton Implementation.
modelopt.torch.quantization.triton.fp4_kernel_hopper	NVFP4 Fake Quantization Triton kernels requiring compute capability >= 8.9 (Hopper+).
modelopt.torch.quantization.triton.fp8_kernel	FP8 Triton Kernel Implementations.

Triton quantization kernels.