up

metascroy · metascroy · commit 5799cfbadb0d · 2024-11-15T11:59:47.000-08:00
diff --git a/extension/llm/export/partitioner_lib.py b/extension/llm/export/partitioner_lib.py
@@ -140,7 +140,7 @@ def _validate_ios_version() -> None:
         minimum_deployment_target=minimum_deployment_target,
         compute_precision=ct.precision(ct.precision.FLOAT16.value),
         # using `ComputeUnit.ALL` can increase the model load time, default to `ComputeUnit.CPU_AND_GPU`
-        compute_unit=ct.ComputeUnit[ct.ComputeUnit.CPU_AND_AE.name.upper()],
+        compute_unit=ct.ComputeUnit[ct.ComputeUnit.CPU_AND_NE.name.upper()],
         model_type=CoreMLBackend.MODEL_TYPE.MODEL,  # pyre-fixme[16]
         op_linear_quantizer_config=op_linear_quantizer_config,
     )

Original file line number	Diff line number	Diff line change
`@@ -140,7 +140,7 @@ def _validate_ios_version() -> None:`
`140`	`140`	`minimum_deployment_target=minimum_deployment_target,`
`141`	`141`	`compute_precision=ct.precision(ct.precision.FLOAT16.value),`
`142`	`142`	# using `ComputeUnit.ALL` can increase the model load time, default to `ComputeUnit.CPU_AND_GPU`
`143`		`- compute_unit=ct.ComputeUnit[ct.ComputeUnit.CPU_AND_AE.name.upper()],`
	`143`	`+ compute_unit=ct.ComputeUnit[ct.ComputeUnit.CPU_AND_NE.name.upper()],`
`144`	`144`	`model_type=CoreMLBackend.MODEL_TYPE.MODEL, # pyre-fixme[16]`
`145`	`145`	`op_linear_quantizer_config=op_linear_quantizer_config,`
`146`	`146`	`)`