Update coreml transformer with Rope and RMSNorm original definitions (#11583)

Summary:
Pull Request resolved: #11583

Keep the module in sync with https://github.com/pytorch/executorch/tree/main/examples/models/llama

Differential Revision: D76468703

Author: YIWENX14

examples/apple/coreml/llama/llama_transformer.py

@@ -12,6 +12,7 @@
 
 import torch
 import torch.nn.functional as F
+from executorch.examples.models.llama.norm import RMSNorm
 
 from executorch.examples.models.llama.rope import (
     hf_apply_rotary_emb,
@@ -64,6 +65,8 @@ class ModelArgs:
     use_scaled_rope: bool = True  # Use scaled RoPE, introduced in llama3.1.
     # Additional Model Metadata needed at runtime
     rope_scale_factor: int = 8
+    high_freq_factor: int = 4
+
     bos_idx: int = 1
     eos_idx: int = 3
     bos_count: int = -1  # i.e., a single EOS is used as BOS
@@ -74,6 +77,12 @@ class ModelArgs:
 
     use_cache_list: bool = True
 
+    use_kv_cache: bool = False
+    enable_dynamic_shape: bool = False
+
+    use_qk_norm: bool = False
+    qk_norm_before_rope: bool = False
+
     def __post_init__(self):
         if self.n_kv_heads is None:
             self.n_kv_heads = self.n_heads
@@ -96,7 +105,7 @@ def __post_init__(self):
             self.head_dim = self.dim // self.n_heads
 
 
-class RMSNorm(torch.nn.Module):
+class CoreMLRMSNorm(torch.nn.Module):
     def __init__(self, dim: int, eps: float = 1e-6):
         """
         Initialize the RMSNorm normalization layer.
@@ -160,10 +169,16 @@ def __init__(self, params: ModelArgs):
         super().__init__()
         self.params = params
         if self.params.use_hf_rope:
-            self.precompute_freqs_cis = hf_precompute_freqs_cis
+            self.precompute_freqs_cis = partial(
+                hf_precompute_freqs_cis,
+                partial_rotary_factor=self.params.partial_rotary_factor,
+            )
         else:
             self.precompute_freqs_cis = partial(
-                precompute_freqs_cis, use_scaled=self.params.use_scaled_rope
+                precompute_freqs_cis,
+                use_scaled=self.params.use_scaled_rope,
+                scale_factor=self.params.rope_scale_factor,
+                high_freq_factor=self.params.high_freq_factor,
             )
         freqs_cos, freqs_sin = self.precompute_freqs_cis(
             self.params.head_dim,
@@ -303,6 +318,14 @@ def __init__(self, args: ModelArgs, layer_id: int, rope: Rope):
 
         self.rope = rope
 
+        self.use_qk_norm = args.use_qk_norm
+        self.qk_norm_before_rope = args.qk_norm_before_rope
+        if self.use_qk_norm:
+            q_norm_dim = self.head_dim
+            k_norm_dim = self.head_dim
+            self.q_norm_fn = RMSNorm(q_norm_dim, eps=args.norm_eps)
+            self.k_norm_fn = RMSNorm(k_norm_dim, eps=args.norm_eps)
+
     def forward(
         self,
         x: torch.Tensor,
@@ -327,6 +350,10 @@ def forward(
         k = k.transpose(1, 2)
         v = v.transpose(1, 2)
 
+        if self.use_qk_norm and not self.qk_norm_before_rope:
+            q = self.q_norm_fn(q)
+            k = self.k_norm_fn(k)
+
         new_k = k
         new_v = v