Add float8_static_activation_float8_weight to user-facing API (#994)

jainapurva · web-flow · commit 09b8b3c3156b · 2024-10-02T14:57:15.000-07:00
diff --git a/torchao/quantization/__init__.py b/torchao/quantization/__init__.py
@@ -46,11 +46,11 @@
     "int4_weight_only",
     "int8_weight_only",
     "uintx_weight_only",
-    "float8_weight_only",
     "fpx_weight_only",
     "LinearActivationQuantizedTensor",
     "to_linear_activation_quantized",
     "to_weight_tensor_with_linear_activation_scale_metadata",
     "float8_weight_only",
-    "float8_dynamic_activation_float8_weight"
+    "float8_dynamic_activation_float8_weight",
+    "float8_static_activation_float8_weight"
 ]
diff --git a/torchao/quantization/quant_api.py b/torchao/quantization/quant_api.py
@@ -93,6 +93,7 @@
     "uintx_weight_only",
     "fpx_weight_only",
     "float8_dynamic_activation_float8_weight",
+    "float8_static_activation_float8_weight",
 ]
 
 from .GPTQ import (

Original file line number	Diff line number	Diff line change
`@@ -93,6 +93,7 @@`
`93`	`93`	`"uintx_weight_only",`
`94`	`94`	`"fpx_weight_only",`
`95`	`95`	`"float8_dynamic_activation_float8_weight",`
	`96`	`+ "float8_static_activation_float8_weight",`
`96`	`97`	`]`
`97`	`98`
`98`	`99`	`from .GPTQ import (`