Method merge

SwissArmyTransformer/sat/mpu/layers.py:478–481

(self, new_weights, new_biases)

            return new_weights, new_biases
    
    def merge(self, new_weights, new_biases):
        self.weight.data.copy_(torch.cat(new_weights, 1))
        if self.bias is not None and self.bias.numel() != 0:
            self.bias.data.copy_(new_biases[0])