From f842c66f0e72fff96222f503571b60caa51bf311 Mon Sep 17 00:00:00 2001
From: Vasiliy Kuznetsov <vkuzo@users.noreply.github.com>
Date: Mon, 16 Sep 2024 16:08:57 -0700
Subject: [PATCH] Update README.md for float8 inference (#896)

---
 torchao/quantization/README.md | 20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

diff --git a/torchao/quantization/README.md b/torchao/quantization/README.md
index 673aae1b5f..433688e44b 100644
--- a/torchao/quantization/README.md
+++ b/torchao/quantization/README.md
@@ -97,7 +97,7 @@ change_linear_weights_to_int4_woqtensors(model)
 
 Note: The quantization error incurred by applying int4 quantization to your model can be fairly significant, so using external techniques like GPTQ may be necessary to obtain a usable model.
 
-#### A16W8 WeightOnly Quantization
+#### A16W8 Int8 WeightOnly Quantization
 
 ```python
 # for torch 2.4+
@@ -109,7 +109,7 @@ from torchao.quantization.quant_api import change_linear_weights_to_int8_woqtens
 change_linear_weights_to_int8_woqtensors(model)
 ```
 
-#### A8W8 Dynamic Quantization
+#### A8W8 Int8 Dynamic Quantization
 
 ```python
 # for torch 2.4+
@@ -121,6 +121,22 @@ from torchao.quantization.quant_api import change_linear_weights_to_int8_dqtenso
 change_linear_weights_to_int8_dqtensors(model)
 ```
 
+#### A16W8 Float8 WeightOnly Quantization
+
+```python
+# for torch 2.5+
+from torchao.quantization import quantize_, float8_weight_only
+quantize_(model, float8_weight_only())
+```
+
+#### A16W8 Float8 Dynamic Quantization with Rowwise Scaling
+
+```python
+# for torch 2.5+
+from torchao.quantization.quant_api import quantize_, PerRow, float8_dynamic_activation_float8_weight
+quantize_(model, float8_dynamic_activation_float8_weight(granularity=PerRow()))
+```
+
 #### A16W6 Floating Point WeightOnly Quantization
 
 ```python