Chirag Jain 50896ec574 Make nvcc threads configurable via environment variable (#885) преди 9 месеца
..
fused_softmax.cpp ed553e9238 Add Megatron attention implementation for benchmarking преди 2 години
scaled_masked_softmax.h ed553e9238 Add Megatron attention implementation for benchmarking преди 2 години
scaled_masked_softmax_cuda.cu ed553e9238 Add Megatron attention implementation for benchmarking преди 2 години
scaled_upper_triang_masked_softmax.h ed553e9238 Add Megatron attention implementation for benchmarking преди 2 години
scaled_upper_triang_masked_softmax_cuda.cu ed553e9238 Add Megatron attention implementation for benchmarking преди 2 години
setup.py 50896ec574 Make nvcc threads configurable via environment variable (#885) преди 9 месеца
type_shim.h ed553e9238 Add Megatron attention implementation for benchmarking преди 2 години