default_per_channel_weight_observer

torch.ao.quantization.observer.default_per_channel_weight_observer

Default per-channel weight observer, usually used on backends where per-channel weight quantization is supported, such as fbgemm.