TransformForMobilePass¶
- class paddle.fluid.contrib.slim.quantization.quantization_pass. TransformForMobilePass [source]
-
-
apply
(
graph
)
apply¶
-
Because paddle-mobile use quantize an dequantize as the names of quantize operator and dequantize operator, the apply function just realize this logic.
- Parameters
-
graph (IrGraph) – the graph will be transformed.
- Returns
-
None
-
apply
(
graph
)