diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py
index ecc09b2145..4e46023a80 100644
--- a/unsloth/models/llama.py
+++ b/unsloth/models/llama.py
@@ -812,11 +812,15 @@ class LlamaRotaryEmbedding(torch.nn.Module):
         inv_freq = 1.0 / (
             self.base ** (torch.arange(0, self.dim, 2, dtype=torch.int64, device="cpu").float() / self.dim)
         )
-        t = torch.arange(self.max_seq_len_cached, device="cpu", dtype=torch.int64).float()
-
-        freqs = torch.outer(t, inv_freq)
-        # Different from paper, but it uses a different permutation in order to obtain the same calculation
+        position_ids = torch.arange(self.max_seq_len_cached, device="cpu", dtype=torch.int64)#.float()
+        inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1)
+        position_ids_expanded = position_ids[:, None, :].float()
+        freqs = (inv_freq_expanded @ position_ids_expanded).transpose(1, 2)
         emb = torch.cat((freqs, freqs), dim=-1)
+
+        # freqs = torch.outer(t, inv_freq)
+        # Different from paper, but it uses a different permutation in order to obtain the same calculation
+        # emb = torch.cat((freqs, freqs), dim=-1)
         self.register_buffer("cos_cached", emb.cos().to(dtype=dtype, device=device, non_blocking=True), persistent=False)
         self.register_buffer("sin_cached", emb.sin().to(dtype=dtype, device=device, non_blocking=True), persistent=False)
     pass