torch.ao.quantization.observer.PlaceholderObserver#
- class torch.ao.quantization.observer.PlaceholderObserver(dtype=torch.float32, custom_op_name='', compute_dtype=None)[源代码]#
Observer that doesn’t do anything and just passes its configuration to the quantized module’s
.from_float().Can be used for quantization to float16 which doesn’t require determining ranges.
- 参数
dtype – Quantized data type
custom_op_name – (temporary) specify this observer for an operator that doesn’t require any observation (Can be used in Graph Mode Passes for special case ops).