Actions: NVIDIA/TensorRT
Actions
Showing runs from all workflows
4,455 workflow runs
4,455 workflow runs
trt.IStreamReader
(as implemented e.g. in polygraphy
) requires higher peak CPU memory and more time than naive python implementation.
Blossom-CI
#6823:
Issue comment #4327 (comment)
created
by
michaelfeil
trt.IStreamReader
(as implemented e.g. in polygraphy
) requires higher peak CPU memory and more time than naive python implementation.
Blossom-CI
#6822:
Issue comment #4327 (comment)
created
by
pranavm-nvidia
trt.IStreamReader
(as implemented e.g. in polygraphy
) requires higher peak CPU memory and more time than naive python implementation.
Blossom-CI
#6821:
Issue comment #4327 (comment)
created
by
michaelfeil
trt.IStreamReader
(as implemented e.g. in polygraphy
) requires higher peak CPU memory and more time than naive python implementation.
Blossom-CI
#6820:
Issue comment #4327 (comment)
created
by
pranavm-nvidia
F.scaled_dot_product_attention
on GPU L4
Blossom-CI
#6815:
Issue comment #4333 (comment)
created
by
ohadravid
F.scaled_dot_product_attention
on GPU L4
Blossom-CI
#6813:
Issue comment #4333 (comment)
created
by
kevinch-nv
F.scaled_dot_product_attention
on GPU L4
Blossom-CI
#6811:
Issue comment #4333 (comment)
created
by
kevinch-nv