Method forward

inference/models/llama.py:32–35

(self, x)

        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)

    def forward(self, x):
        output = torch.empty_like(x)
        awq_inference_engine.layernorm_forward_cuda(x, self.weight, output, self.eps)
        return output


def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0):