Default placeholder observer, usually used for quantization to torch.float16.
上一页
torch.ao.quantization.observer.default_per_channel_weight_observer
下一页
torch.ao.quantization.observer.default_weight_observer