fused_per_channel_wt_fake_quant_range_neg_127_to_127

torch.ao.quantization.fake_quantize.fused_per_channel_wt_fake_quant_range_neg_127_to_127

Fused version of default_per_channel_weight_fake_quant, with the 8-bit values restricted to [-127, +127], excluding -128.