Unverified Commit 1568c6ed authored by Kirthi Shankar Sivamani's avatar Kirthi Shankar Sivamani Committed by GitHub
Browse files

fix documentation for default FP8 recipe format (#6)


Signed-off-by: default avatarKirthi Shankar Sivamani <ksivamani@nvidia.com>
Signed-off-by: default avatarKirthi Shankar Sivamani <ksivamani@nvidia.com>
parent 2db83763
...@@ -63,7 +63,7 @@ class DelayedScaling: ...@@ -63,7 +63,7 @@ class DelayedScaling:
Margin for the scaling factor computation. Margin for the scaling factor computation.
interval : int, default = 1 interval : int, default = 1
Controls how often the scaling factor is recomputed. Controls how often the scaling factor is recomputed.
fp8_format : {Format.E4M3, Format.HYBRID}, default = Format.E4M3 fp8_format : {Format.E4M3, Format.HYBRID}, default = Format.HYBRID
Controls the FP8 data format used during forward and backward Controls the FP8 data format used during forward and backward
pass. pass.
amax_history_len : int, default = 1 amax_history_len : int, default = 1
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment