Update autoround RTN example to tuning (#2558)

mengniwang95 · web-flow · commit 7536f0373c87 · 2026-04-02T13:03:30.000-04:00
SUMMARY:

Update autoround RTN example to tuning

Signed-off-by: Wang, Mengni &lt;mengni.wang@intel.com&gt;
diff --git a/examples/autoround/quantization_w8a8_fp8/llama3.1_block_quant_example.py b/examples/autoround/quantization_w8a8_fp8/llama3.1_block_quant_example.py
@@ -25,7 +25,7 @@
 # Configure the quantization algorithm to run.
 # NOTE: AutoRoundModifier with iters=0 is equivalent to RTN
 recipe = AutoRoundModifier(
-    targets="Linear", scheme="FP8_BLOCK", ignore=["lm_head"], iters=0
+    targets="Linear", scheme="FP8_BLOCK", ignore=["lm_head"], iters=200
 )
 
 

Original file line number	Diff line number	Diff line change
`@@ -25,7 +25,7 @@`
`25`	`25`	`# Configure the quantization algorithm to run.`
`26`	`26`	`# NOTE: AutoRoundModifier with iters=0 is equivalent to RTN`
`27`	`27`	`recipe = AutoRoundModifier(`
`28`		`- targets="Linear", scheme="FP8_BLOCK", ignore=["lm_head"], iters=0`
	`28`	`+ targets="Linear", scheme="FP8_BLOCK", ignore=["lm_head"], iters=200`
`29`	`29`	`)`
`30`	`30`
`31`	`31`